JOHN WILEY & SONS, LTD., THE ATRIUM, SOUTHERN GATE, CHICHESTER P019 8SQ, UK

*** PROOF OF YOUR ARTICLE ATTACHED, PLEASE READ CAREFULLY ***

After receipt of your corrections your article will be published initially within the online version of the journal.

PLEASE NOTE THAT THE PROMPT RETURN OF YOUR PROOF CORRECTIONS WILLENSURE THAT THERE ARE NO UNNECESSARY DELAYS IN THE PUBLICATION OF

YOUR ARTICLE

READ PROOFS CAREFULLY

ONCE PUBLISHED ONLINE OR IN PRINT IT IS NOT POSSIBLE TO MAKE ANY FURTHERCORRECTIONS TO YOUR ARTICLE

§ This will be your only chance to correct your proof§ Please note that the volume and page numbers shown on the proofs are for position only

ANSWER ALL QUERIES ON PROOFS (Queries are attached as the last page of your proof.)

§ List all corrections and send back via e-mail to the production contact as detailed in the covering e-mail,or mark all corrections directly on the proofs and send the scanned copy via e-mail. Please do not sendcorrections by fax or post

CHECK FIGURES AND TABLES CAREFULLY

§ Check sizes, numbering, and orientation of figures§ All images in the PDF are downsampled (reduced to lower resolution and file size) to facilitate Internet

delivery. These images will appear at higher resolution and sharpness in the printed article§ Review figure legends to ensure that they are complete§ Check all tables. Review layout, titles, and footnotes

COMPLETE COPYRIGHT TRANSFER AGREEMENT (CTA) if you have not already signed one

§ Please send a scanned signed copy with your proofs by e-mail. Your article cannot be publishedunless we have received the signed CTA

OFFPRINTS

§ 25 complimentary offprints of your article will be dispatched on publication. Please ensure that thecorrespondence address on your proofs is correct for dispatch of the offprints. If your delivery addresshas changed, please inform the production contact for the journal – details in the covering e-mail.Please allow six weeks for delivery.

Additional reprint and journal issue purchases

§ Should you wish to purchase a minimum of 100 copies of your article, please visithttp://www3.interscience.wiley.com/aboutus/contact_reprint_sales.html

§ To acquire the PDF file of your article or to purchase reprints in smaller quantities, please visithttp://www3.interscience.wiley.com/aboutus/ppv-articleselect.html. Restrictions apply to the use ofreprints and PDF files – if you have a specific query, please contact permreq@wiley.co.uk.Corresponding authors are invited to inform their co-authors of the reprint options available

§ To purchase a copy of the issue in which your article appears, please contact cs-journals@wiley.co.ukupon publication, quoting the article and volume/issue details

§ Please note that regardless of the form in which they are acquired, reprints should not be resold, norfurther disseminated in electronic or print form, nor deployed in part or in whole in any marketing,promotional or educational contexts without authorization from Wiley. Permissions requests should bedirected to mailto: permreq@wiley.co.uk

Research Article 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

Received: 18 April 2008, Revised: 12 February 2009, Accepted: 4 April 2009, Published online in Wiley InterScience: 00 Month 2009

63

64

(www.interscience.wiley.com) DOI:10.1002/nbm.1408

65

66

67

68

69

70

71

72

73

74

75

76

77

S

Classification of prostatic diseases by meansof multivariate analysis on in vivo proton MRSIand DCE-MRI dataMariacristina Valerioa, Valeria Panebiancob, Alessandro Sciarrac,Marcello Osimanib, Stefano Salsicciac, Lorena Casciania,Alessandro Giulianid, Mariano Bizzarrie, Franco Di Silverioc,Roberto Passariellob and Filippo Contia*

78

Multivariate analysis has been applied on proton ma

NMR Biom

79

80

81

82

83

84

85

86

87

88

89

ED PRO

OF

gnetic resonance spectroscopic imaging (1H-MRSI) and dynamiccontrast enhanced MRI (DCE-MRI) data of patients with different prostatic diseases such as phlogosis, fibrosis andadenocarcinoma. Multivariate analysis offers a global view of the entire range of information coming from both theimaging and spectroscopic side of NMR technology, leading to an integrated picture of the system relying upon theentire metabolic and dynamic profile of the studied samples. In this study, we show how this approach, applied to1H-MRSI/DCE-MRI results, allows us to differentiate among the various prostatic diseases in a non-invasive way with a100% accuracy. These findings suggest that multivariate analysis of 1H-MRSI/DCE-MRI can significantly improve thediagnostic accuracy for these pathological entities. From a more theoretical point of view, the complementation ofsingle biomarker approach with an integrated picture of the entire metabolic and dynamic profile allows for a morerealistic appreciation of pathological entities. Copyright � 2009 John Wiley & Sons, Ltd.

Keywords: cancer; differential diagnosis; dynamic contrast enhanced MRI; medical imaging; metabolomics; multivariateanalysis; prostatic diseases; proton magnetic resonance spectroscopic imaging

90

91

T

* Correspondence to: F. Conti, Department of Chemistry, University of Rome ‘LaSapienza’, Rome, Italy.E-mail: filippo.conti@uniroma1.it

a M. Valerio, L. Casciani, F. Conti

Department of Chemistry, University of Rome ‘La Sapienza’, Rome, Italy

b V. Panebianco, M. Osimani, R. Passariello

Department of Radiological Sciences, University of Rome ‘La Sapienza’, Rome,

Italy

c A. Sciarra, S. Salsiccia, F. Di Silverio

Department of Urology, University of Rome ‘La Sapienza’, Rome, Italy

d A. Giuliani

Department of Environment and Health, Istituto Superiore di Sanita, Rome,

Italy

e M. Bizzarri

Department of Experimental Medicine, University of Rome ‘La Sapienza’,

Rome, Italy

Abbreviations used: AC, adenocarcinoma; BPH, benign prostatic hyperpla-

sia; Cho, choline; Cit, citrate; CO, control; Cr, creatine; DCE-MRI, dynamic

contrast-enhanced magnetic resonance imaging; FB, fibrosis; 1H-MRSI, proton

magnetic resonance imaging; LVs, latent variables; OT, onset time; PCA,

principal component analysis; PCs, principal components; PE, peak enhance-

ment; PL, phlogosis; PLS-DA, partial least square discriminant analysis; PrC,

prostate cancer; PSA, prostate-specific antigen; ROI, region-of-interest; SI-T,

signal intensity-time; TRUS, transrectal ultrasonography; TTP, time to peak;

VOI, volume of interest. 1

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

UNCORRECINTRODUCTION

Differential radiological diagnosis of both benign and malignantprostatic diseases, such as benign prostatic hyperplasia (BPH),prostatitis and prostate cancer (PrC) is often difficult despite theirepidemiological and clinical relevance (1,2). The non-invasivedifferential diagnosis of prostate cancer, BPH, prostatisits andnormal tissue is of utmost importance for cancer staging and forfollow-up after therapy (3).The serum prostate-specific antigen (PSA) has been identified

as a sensitive biological marker for prostate cancer diagnosis inrecent times. However, the PSA levels, although significantlylinked to prostate cancer, are devoid of any discriminatory powerfor deciding among different prostatic diseases (5). In fact,elevated serum PSA levels can be caused by benign conditionsthat are mainly prevalent in older men (6). Till date, histologicalanalysis of biopsy specimens has been the only reliableprocedure to distinguish normal, benign and malignant prostatictissues. However, due to the heterogeneous and frequentmultifocal nature of prostate cancer, biopsy methods may notinclude an adequate specimen sampling of the prostate. In fact,clinical studies have indicated that the usual systematic sextantbiopsy technique shows a positive predictive value of only 30%for detection of prostate cancer (7). The use of novel biopsyschemes significantly increases the diagnostic yield of prostatebiopsy in finding the malignant disease (8–10), but still does notreach a fully satisfactory accuracy of cancer mapping within theprostate (11).

ed. 2009; 22: 1–11 Copyright � 2009 John Wiley & Sons, Ltd.

119

120

T

M. VALERIO ET AL.

2

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

UNCORREC

Improvements in prostate imaging provide more accuratemapping of cancer allowing for biopsy optimized plans. Needlebiopsy guided by transrectal ultrasonography (TRUS) is the mostcommonly used method for histological diagnosis of prostatecancer. On the other hand, this technique is limited by the lack ofaccuracy in the estimation of cancer tissue extension (12).High-resolution endo-rectal/pelvic phased array MRI has

demonstrated an enhanced sensitivity compared to clinicaldata, systematic biopsy, TRUS and MRI when considered alone,but it provides low specificity in detection and localization ofprostate cancer due to other benign pathologies which cause lowsignal intensity on T2-weighted images similar to that of prostatecancer (13–15). Moreover, this conventional imaging methodsometimes cannot distinguish accurately between healthy andmalignant tissues subsequent to resection or treatments forprostate cancer because of induced changes in tissue structure(16). The ability to identify prostate tissue types can besignificantly improved by the combined use of MRI and magneticresonance spectroscopic imaging (MRSI) (17,18). Specifically,proton MRSI (1H-MRSI) allows the recognition of some relevantmetabolites like citrate (Cit), choline (Cho) and creatine (Cr),endowed with the discrimination power for different prostatediseases. In particular, the ratio choline-plus-creatine to citrate([ChoþCr]/Cit) has been widely investigated in the differentialdiagnosis of prostate diseases. The rationale at the basis of theuse of this index stems from the observed increase ofCho-containing compounds and on the reduction or absenceof Cit in PrC compared to BPH and the surrounding healthyperipheral zone tissue. However, in some cases, it has beenreported that glandular BPH and normal, healthy peripheral zonetissues display similar Cit levels; on the other hand, stromal BPHregions can show reduced Cit levels similar to those observed inperipheral zone cancers (19–26). In a recent paper, Li et al. (27)analysed the spectral differences between PrC and BPH,evaluating the (Choþ Cr)/Cit and Cho/Cr ratios measured ineach voxel with proven-biopsy cancer or BPH. The specificity,sensitivity and accuracy for the discriminant function were 98.6,85.7, 92.9%, respectively. Finally, the addition of metabolicinformation provided by MRSI to morphologic informationprovided an enhanced specificity up to 95% for localizing cancerand has become important for detecting the extent of cancerwithin the prostate and its aggressiveness (17,18,28). Despite thevery high accuracy of the MRI/MRSI combined approach in thedetection of prostate cancers, this mixed technique is not veryaccurate in identifying cancers (only 30% are found) in both thecentral gland (29) and small (<0.5 cm3) cancers within theperipheral zone (21,30).Recent studies have shown that accuracy can be improved by

performing MRI/MRSI at higher magnetic field strengths andthrough the addition of dynamic contrast enhanced MRI(DCE-MRI) (31–33).DCE-MRI is known to be a powerful tool for visualizing the

vascularity of prostate gland tumour and for providing additionalinformation useful for both the detection and the staging ofprostate cancer (31,34). Dynamic MRI was demonstrated to beable to differentiate cancer from normal prostatic tissue (35,36),an earlier and stronger enhancement in PrC versus normal tissuewas found in these studies. DCE-MRI was also demonstrated todifferentiate the cancer tissue from benign lesions (37,38).Recently, Ren et al. (39) reported that, based on T2-weightedimaging, DCE-MRI curves can discriminate PrC and BPH with asensitivity, specificity and accuracy of 79.31, 66.67 and 74%,

www.interscience.wiley.com/journal/nbm Copyright � 200

ED PRO

OFS

respectively. In particular, PrC showed stronger enhancementwith an earlier peak time, higher enhancement and enhance-ment rate than those of BPH.In this study, we developed a computational approach to the

analysis of 1H-MRSI and DCE-MRI combined results, that wasdemonstrated to outperform both the classical ([ChoþCr]/Cit)ratio and purely image-based parameters in the differentialdiagnosis of prostatic diseases.Our data set was composed of 11 prostatic tissue samples

coming from healthy control subjects, and of 40 pathologic tissuesamples coming from patients affected by one of the followingpathologies: (a) phlogosis, (b) fibrosis or (c) adenocarcinoma.The discrimination between healthy and disease samples and thedifferential diagnosis of the various prostatic diseases was theaim of the present study. Given the paucity of the data set, ourwork has mainly methodological value, giving a proof-of-conceptto the feasibility of an integrative systemic view,made possible bymultivariate approach, to the MRI/MRSI non-invasive differentialdiagnosis of prostatic diseases.We adopted a data analysis strategy that mixed unsupervised

and supervised approaches. Unsupervised is the general headingof techniques such as principal component analysis (PCA) andcluster analysis, whose final result is not guided by themaximization of ‘externally’ imposed classification goals, likethe discrimination between different classes of disease orplacebo and drug treated patients. The goal of unsupervisedalgorithms is to maximize some purely syntactical internalfeatures of the data set at hand, like the obtaining of the mostfaithful projection of an initially high dimensional data set withthe least number of axes (PCA), or the allocation of the statisticalunits to classes that are the most internally compact andseparated (cluster analysis) (40). Unsupervised methods allow foran unbiased (not driven by the goal of diagnosis) description ofthe natural correlation structure present in the data. In contrast,supervised methods that we adopted in our strategy, like partialleast squares (PLS) or discriminant analysis (DA), have the goal ofmaximizing an ‘externally imposed task’, such as the separation oftwo a priori classes (like healthy/disease) inside a given data field.In fact, supervised methods are driven by a specific goal, externalto the intrinsic nature of the collected data, assuring the ‘bestpossible discrimination’ of the classes, at the expense of theappreciation of the natural correlations present in the data, whichwe exploited with the unsupervised approach.We decided to complement the unsupervised and supervised

approaches, with a classical strategy of data analysis (41) so as toget the maximal global efficiency of the model on bothdescriptive and diagnostic sides. The initial unsupervisedextraction of principal components from the original data setallows for noise filtering of the data, maintaining only thecorrelated (and thus more reliable) portion of information andpermitting a biological interpretation of the obtained results.Subsequently, the use of the extracted principal components asinitial variables for the supervised portion of the procedure,avoids possible inconsistencies coming from the regressors’mutual collinearity components which are orthogonally con-structed. However, besides statistical subtleties, the importantthing to stress is that the adopted strategy allows for all theinformation embedded in our data to be potentially exploited forthe task of classification.All in all, the obtained results allowed for both an efficient

discrimination of the different diseases and for a biologicallysound general picture of the system at hand.

9 John Wiley & Sons, Ltd. NMR Biomed. 2009; 22: 1–11

119

120

T

METABOLOMIC ANALYSIS OF PROSTATIC DISEASES

3

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

UNCORREC

METHODS

Patient population

In this study, we retrospectively reviewed a total of 51 prostateMR examinations, including morphologic imaging, 1H-MRSI andDCE-MRI protocols, carried out from June 2007 to February 2008.The MR investigations were performed before TRUS-guidedneedle biopsy. The mean time interval between MR examinationand TRUS-guided biopsy was 7� 3 days. For each patient,medical histories including digital rectal examination, serum PSAlevel and a confirmed biopsy report were obtained. Our studypopulation can be subdivided into four groups based on the kindof prostate disease clinically and histopathologically diagnosed.The first group consisted of 13 patients with biopsy-proven PrCwho subsequently underwent radical retropubic prostatectomyperformed within 3 weeks (mean¼ 9� 4 days). Histolopatholo-gical examination of the radical prostatectomy specimenrevealed a mean Gleason score of 7� 2 (range¼ 6–10) andthe absence of BPH nodules. In the second group, we included 14patients with biopsy-proven BPH who subsequently underwenttrans-urethral resection. Pathologic assessment of the resectedtissue confirmed TRUS results. The mean period between MRexaminations and trans-urethral resection was 12� 7 days. Thethird group was composed of 13 patients with a biopsy-provenfibrosis performed for a TRUS suspicious prostate nodule. Acontrol group of 11 patients with no pathological findings of PrC,BPH and fibrosis was also evaluated. In this group, a prostateTRUS-biopsy was performed following a suspicious clinicalexamination and/or rising serum PSA level (mean¼4.9� 7.0 ng/mL; median¼ 1.7 ng/mL; range¼ 0.5–18 ng/mL).Furthermore, digital rectal examination did not reveal anyprostate abnormalities.The following exclusion criteria were used in the selection of

patient population: clinical diagnosis of acute prostatitis,coexistent clinically proven cancer, hormonal therapies (includingfive a reductase inhibitors), radiotherapy, chemotherapy,previous prostate surgery.The 51 included patients had a mean age of 65 (range 48–75

years); age was checked for its possible confounding effect on theMR-based parameters, without finding any significant correlationwith the descriptors under study (data not shown).This study was approved by the local ethics committee and

signed informed consent was obtained from all patients.

MRI, 1H-MRSI and DCE-MRI data acquisition and processing

Acquisition of imaging data

All examinations were performed on a commercially available1.5 T scanner (Magnetom Avanto, Siemens Medical Solutions,Erlangen, Germany), equipped with surface phased-array (BodyMatrix, Siemens Medical Solutions) and endo-rectal coil (e-Coil,Medrad, Pittsburgh, PA, USA, combined with Endo-Interface,Siemens Medical Solutions). The balloon-mounted disposableendo-rectal coil was first lubricated with a local anaesthetic geland then placed while the patient was in the left lateral decubitusposition. Then the patient was turned supine and the balloon wasinflated with up to 70mL of room air based on patient’s tolerance.Before scanning, 20mg butyl scopolamine (Buscopan, Boehrin-ger, Ingelheim, Germany) was injected to suppress peristalsis.First, localizer images in the sagittal, axial and coronal planes

were obtained to ensure endo-rectal coil position and to select

NMR Biomed. 2009; 22: 1–11 Copyright � 2009 John Wiley & S

ED PRO

OFS

locations for the transverse images. Following this, T2-weightedimages in the three orthogonal planes were acquired providingcoverage of the entire prostate using turbo spin-echo (TSE)sequences (TR¼ 5190ms; TE¼ 95ms; flip angle¼ 1508;average¼ 3; FOV read¼ 256mm; FOV phase¼ 100; slice thick-ness¼ 3mm; interslice gap¼ 0; matrix size¼ 512� 512; phaseresolution¼ 100%; bandwidth¼ 130Hz; scan time¼ 3.40min).

1H-MRSI data were acquired by two skilled radiologists after afirst review of morphological images to localize suspicious areasin the prostate, which were subsequently used to position thespectroscopic acquisition volume. In patients with no MRmorphological evidence of changes, the volumes of interest(VOIs) were centred on each prostate gland emi-portion (left andright). The VOI to be studied with spectroscopy was selected insuch a way, as to maximize the coverage of the prostate whileminimizing the inclusion of surrounding structures (muscles, fat,rectal air and urine). 1H-MRSI was performed using a section-selected box drawn closely around the prostate fossa and apoint-resolved spectroscopic sequence was obtained by using a3D chemical shift imaging (CSI) sequence (FOV¼ 50�50� 50mm3; VOI¼ 30� 30� 30mm3; TR¼ 700ms; TE¼ 120ms;ms; flip angle¼ 908; interpolation¼ 16; vector size¼ 512;TA¼ 11.50min; delta frequency¼�1.80 ppm; average¼ 6;filter¼Hamming) (42).DCE-MRI images were acquired using 3D FLASH T1-weighted

spoiled gradient-echo sequence (TR¼ 2.44ms; TE¼ 0.9ms; flipangle¼ 308; average¼ 1; thickness¼ 4mm; interslice gap¼ 0;slice number¼ 12; matrix size¼ 256� 256; phase resolution¼100%; bandwidth¼ 120Hz; TA¼ 4.40min) performing 90measurements in rapid succession, immediately after thecompletion of an intravenous bolus injection of 0.1mmol ofgadopentetate dimeglumine (Multihance, Bracco Spa, Milano,Italy). Contrast liquid was administered with a power injector(Spectris; Medrad) at 2.5mL/s and was followed by a 15-mL salineflush. The 3D volume was acquired with the same positioningangle and centre as the transverse T2-weighted sequencecovering the entire prostate fossa and the periurethral-perianastomotic region. Relative gadolinium chelate concen-tration curves were calculated in order to derive the threedynamic DCE-MRI parameters: onset time, time to peak and peakenhancement (PE).

Processing and analysis of imaging data

MR images were analysed in consensus by two radiologists with 5and 9 years of experience in uro-genital MRI. They were unawareof serum PSA levels and TRUS-biopsy results. T2-weighted imageswere excluded from retrospective reviewing and the radiologists’attempts were focused only on spectroscopy and DCE-MRIfindings.An operator-independent standard post-processing protocol

was applied to the MR spectroscopic imaging data. These datawere acquired as 16� 8� 8 phase-encoded spectral arrays (1024voxels) with a nominal spatial resolution >0.3 cm3 before Fouriertransformation in the spatial dimensions. After Fourier trans-formation, zero- and first-order phase correction and automatedbaseline correction (polynomial of 6th order), a frequencydomain curve fitting was used subsequently for quantificationwith the assumption of Gaussian line shapes, by using thestandard Syngo Spectroscopic Evaluation software package(Siemens), provided with the MR imaging system (43). Goodnessof fit of the obtained parameter by means of the classical

ons, Ltd. www.interscience.wiley.com/journal/nbm

119

120

T

M. VALERIO ET AL.

4

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

UNCORREC

Gaussian distribution hypothesis was assessed by subtracting theprocessed spectrum from the fitted one, and checking that onlysignals indistinguishable from the baseline noise remained at thementioned ppm in the residual curves. We used areas under thecurve to compute Cho (3.2 ppm), Cr (3.0 ppm) and Cit (2.6 ppm)values. The average post-processing duration was 25–30min foreach data set. Cho, Cr and Cit peak areas were evaluated for all1H-MRSI voxels not contaminated by inadequately suppressedwater or lipids, and did not contain mixed tissues from theurethra, seminal vesicles, ejaculatory ducts, and bladder andrectal wall. In addition, only voxels in which the choline/creatine/citrate peaks were detectable with a signal-to-noise ratio of 3:1were assessed (44).The dynamic MR post-processing procedure lasted 10min for

each patient. Functional dynamic imaging parameters wereestimated from the gadolinium curve using the procedure ofFutterer et al.(19). The edge and the contour characteristics of thelesions were defined using the same sections on which theregion-of-interest (ROI) analyses were performed.A group of three ROIs were drawn independently by the

radiologists, and differences in the measurements were assessedby consensus. The selected ROIs referred to three distinct areas:(1) pelvic muscle (acting as low baseline ROI), (2) most-enhancingareas within the main PrC foci, BPH nodules, fibrosis nodules or inthese regions with suspect spectroscopic voxels and (3) iliacvessel (acting as high baseline ROI). In particular, suspect regionswere identified based on higher enhancing values on DCE-MRIimages (qualitative method). Correspondingly, normal tissue wasidentified as the one having homogenous enhancing regions.When multiple suspicious areas were identified, the signalintensity–time (SI–T) records of the most enhancing lesion wereconsidered as significant values for subsequent SI–T analyses. Thefollowing parameters were set to describe the SI time curve:onset time, time to peak and PE. We determined theenhancement onset time for the data sets by averaging (during90 measurements) the intensity across the slices and using thelast point before the averaged signal increased 2.5 standarddeviations (SDs) above the running baseline average.

TRUS-biopsy evaluation

TRUS guided biopsies were performed using a biplanar 7.5MHzfrequency probe according to a 12-core biopsy scheme (standardsextant scheme, plus laterally directed samples of the prostateapex (two cores), middle (two cores) and base (two cores)) (10).No samples in the transitional zone of the prostate wereobtained. The operator was blinded from 1H-MRSI and DCE-MRIresults. All biopsy specimens were obtained under TRUSguidance using an 18-gauge needle loaded in an automaticspring action biopsy device, and were fixed overnight in asolution of 10% neutral buffered formalin. The operator evaluatedthe distance from the prostate apex and basis and the distancefrom the urethra from which the biopsy specimen was drawn, inorder to provide a method of comparison with T2-weighted MRimages as reference. MRI films were interpreted independently bya third radiologist with 6 years of experience in uro-genitalradiology, who had no knowledge of 1H-MRSI and DCE-MRIresults and of the final diagnosis. Each axial T2-weighted imagewas branched in 12 radial triangles with apex orientation on theurethra in a clockwise order. Distance from the prostate apex andbasis and the distance from the centre were noted by the

www.interscience.wiley.com/journal/nbm Copyright � 200

ED PRO

OFS

Radiologist. For each sample, all abnormalities were examined inconsensus by the biopsy operator and radiologist.

Statistical analysis

Raw data matrix structure

Multivariate analysis was applied to the data set constituted bythe 1H-MRSI spectral and dynamic DCE-MRI parametersmeasured on healthy prostatic tissues of 11 control (CO) as wellas on the pathologic tissues of 40 patients who had one of thefollowing pathologies: (1) a phlogosis (PL, n¼ 14); (2) a fibrosis(FB, n¼ 13) or (3) adenocarcinoma (AC, n¼ 13). This produced araw data set constituted by a matrix having as rows (statisticalunits) 51 patients and as column (variables) the 6 values relativeto choline, creatine, citrate (all these three variables are expressedin terms of the area of the relative peak), onset time, time to peak(these dynamic descriptors are expressed in seconds) and PEparameters (c� (mmol/kg contrast agent)) obtained from1H-MRSI/DCE-MRI measurements. The original 51 units in thepopulation were separated into two sets: a training and avalidation (test) set. The training set was made of 45 subjects,while 6 subjects (2 PL, 2FB, 2AC) composed the test set. Themodel was built upon the 45 subjects in the training set and thenwas checked on its ability to correctly classify the set of sixpatients. This procedure known as cross-validation allows for thetesting of the generalization ability of the proposed model,outside the range of the specific data set it comes from. Thiseliminates overfitting and chance correlation problems, which isparticularly important in this case, where the paucity of datagreatly increases the risk of apparent correlations (41).

Principal component analysis

PCA is a projection method used for exploiting the informationembedded in multidimensional data sets (40). The data arereduced to a few latent variables (LVs) (or principal components)collecting the information implicit in the original variables’correlation structure. The extracted components (PCs) are eachorthogonal and ordered in terms of the percentage of explainedvariation, with the first components collecting the ‘signal’(correlated) portion of information, while minor componentscan be considered as ‘noise’ components. From an algebraicpoint of view, each component is a weighted summationcomputed across the original variables in the form of:PC¼ aX1þbX2þ cX3. where X1, X2 and X3 are the measuredfeatures and a, b, c numerical constants. Each statistical unit isassigned a score relative to each extracted component, while thecorrelation coefficient between each original variable andextracted components (loading) allows to give a meaning tothe PCs.

Partial least square discriminant analysis

While PCA is an unsupervised technique in which each variableenters with the same role of description of the data set andthe solution is driven by the maximal parsimony principle alone(maximal amount of explained variation with the minimumnumber of components); on the other hand, both PLS and DA aresupervised techniques in which the analysed variables pertain totwo classes: the ‘diagnosis’ (dependent, Y) and ‘symptoms’(independent, X) variables. The goal of both PLS and DA is to findthe linear combination of X variables that explains the Y

9 John Wiley & Sons, Ltd. NMR Biomed. 2009; 22: 1–11

119

120

METABOLOMIC ANALYSIS OF PROSTATIC DISEASES 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

variable(s) better. In the case of DA, this goal is achieved by theconstruction of a set of weights multiplying each X variable so asto build metrics in which the errors of assignment of eachstatistical unit to the correct Y class is minimized. PLS, on theother hand, works by the generation of mutually orthogonallinear combinations of X variables maximally correlated with Ycounterparts. PLS-DA was used to build and test a supervisedmodel that could predict the pathology of a patient based on itsspectral and dynamic data. ‘Leave-one-out’ cross validationmethod was used to validate the model and to select theappropriate number of LV. The identification and the removal ofoutliers were performed by using the Q and T2 statistics (45).

Range restriction effect check

The presence of outliers is known to deeply influence thecorrelation analyses by means of the so called range restrictioneffect (46). This effect has to do with the fact that outliers(extreme values observations) acquire a disproportionate weightin the computation of any least-squares- based model. In fact,optimizing the fit of an extreme point (outlier) has a much greaterinfluence in the sum of errors (to be minimized by the system)than the best fitting of an average unit. This statistical effect couldgive rise to biased descriptions of the studied data sets. For thisreason, we decided to perform two independent analyses: thefirst one on the complete data set (45 units) and the second oneon the same data sets depurated by four possible outliers (41units). We considered as outliers the units being located at morethan 3 SD units from the centroid of the data cloud.

T

5

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

UNCORREC

RESULTS

Multivariate analysis of 1H-MRSI and DCE-MRI data

Individual and mean values of spectral 1H-MRSI and dynamicDCE-MRI parameters measured on healthy prostatic tissues ofcontrol as well as on the pathologic tissues of patients withphlogosis, fibrosis or adenocarcinoma are shown in Fig. 1. Thedata indicate the presence of a substantial overlap amonggroups, even if we must take into account the fact thatthe relatively high variances of the considered variables is partlydue to the coil sensitivity profiles. Nevertheless, correction factorsare not routinely used in such kind of analyses. This figure allowsto immediately perceive the huge advancement obtained byexplicitly taking into account the fact that the same statistical unit(patient) is simultaneously defined over different variables. In fact,the possibility of distinction relies on the correlation betweenvariables instead of considering each descriptor in isolation fromall the others (47).Before entering the actual statistical strategy with the explicit

consideration of the correlation structure of the measuredfeatures, we adopted the classical approach applied in thiscategory of problems by submitting the descriptors most widelyused for differential diagnosis to DA. The discriminant functionwas computed over the (ChoþCr)/Cit ratio and OT, TTP and PE.The specificity, sensitivity and accuracy are reported in Table 1 fordifferences between class comparisons. Although all theconsidered descriptors displayed a statistically significantdiscriminative power, there are some specific inter-groupcomparisons that were markedly sub-standard. In contrast, wewill show how the multivariate approach not only reaches amaximal accuracy on the entire set of comparisons, but how it is

NMR Biomed. 2009; 22: 1–11 Copyright � 2009 John Wiley & S

ED PRO

OFS

also able to correctly predict the test set statistical units, thusdemonstrating a predictive ability outside the realm of thesamples used to build the model.Data analysis begun with the computation of PCA to identify

any clustering of data related to the types of pathology in anunsupervised manner, while PLS-DAwas subsequently applied tobuild a classification model to predict the clinical outcome of apatient based on its spectral and dynamic data and usable evenfor independently analysed patients (test set) (45).PCA applied to the original 45 units/6 variables data set

([ChoþCr]/Cit ratio is a derived variable and its information isimplicit in the original descriptors), gave rise to a four componentsolution explaining about 87% of the total variability in thesystem. In Table 2 the variance explained by each component isreported. To compare the controls and patients (as a whole), at-test was applied to the component scores, highlightingsignificant differences between the two groups on PC1 andPC3 (see Table 2). This can be appreciated in Fig. 2, where thecomponent score plot is shown. A linear discriminant analysisapplied to this space allowed for a clear separation of the twogroups (Fisher’s exact test, p< 0.0001 on the classificationmatrix).The t-test on PCs is then repeated to compare all the specific

pairs of control and pathology groups (see Table 2). As a result, PLand AC patients showed significant differences compared tocontrols on both PC1 and PC3 (Fisher’s exact test p< 0.0001),whereas the FB group differed from controls as for PC2 and PC3(Fisher’s exact test p< 0.0001). Figures 2 and 3 display theabove-mentioned differences in various component planes.We subsequently analysed the data by performing a t-test on

PCs to compare a single pathology (see Table 2). PC1 wasresponsible for the discrimination among pathologies andpairwise discriminations are also observed on PC2 and PC4(Fisher’s exact test p< 0.0001) for PL versus FB as well as PL versusAC groups (see Fig. 4). The above results indicate that even byusing this unsupervised method of analysis, with only a posterioricomputation of the statistical significance over the extractedcomponents, there is a clear separation between controls andpathologic groups, controls and single-pathologic groups andamong patients with different types of pathology.After having checked for the general robustness of the

correlation structure, we considered the possibility of usingmultivariate methods to classify patients not used for modelbuilding. This means shifting from a purely descriptive to apractical diagnostic use of the technique that by definition mustbe effective in predicting the diagnosis of samples not explicitlytaken into account for statistical model building. For this goal, webuilt a PLS-DA model based on 45 patients data set. This method,at odds with PCA, is a supervised procedure explicitly driven bythe optimization of discrimination power of the model. Thefeasibility of the proposed method as a routine diagnosticprocedure depends on the successful classification of the test set.The PLS-DA model generated three LVs which explain 70% of

the X-variance (spectral and dynamic data) and 50% of theY-variance (which represent the membership class). Table 3summarizes the features of this model. A clear separation amongthe classes was found in the first and second components (Fig. 5)thus confirming the unsupervised approach (PCA). This modelwas used to assess the predictive capabilities for six otherpatients with unknown pathology. The model predicted that twoof the unknown samples were from the PL group, two from the FBand two from the AC group. The actual identities of the samplescoincided with the predicted ones (i.e. the sensitivity and

ons, Ltd. www.interscience.wiley.com/journal/nbm

119

120

CORRECTE

D PRO

OFS

Figure 1. Plot of individual (open circle) and mean (filled circle) values of spectral 1H-MRSI (area) and dynamic DCE-MRI (onset time and time to peak in

seconds; peak enhancement in c� (mmol/kg contrast agent) parameters measured on healthy prostatic tissue of normal control as well as on thepathologic tissue of patients with phlogosis, fibrosis or adenocarcinoma.

M. VALERIO ET AL.

6

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

UN

specificity for the PLS model based on pathology are both 100%),thus providing a validation of the PLS model as a diagnostic tool(see Table 4).Having proven the efficacy of the model as a predictor, we

decided to gomore in depth into the nature of the metabolic anddynamic biomarkers which permit discrimination. To this aim, wederived a PLS model using a reduced data set by eliminatingoutliers that could bias the solution by their excessive weight incomputing the correlation. The outliers were removed by meansof the procedure previously outlined in the Method section. Theremoval of outliers provided a final 41 statistical units and a rawdata matrix of six variables. The loading plot relative to thediscriminating variables is shown superimposed over a score plotin Fig. 6. This representation allows to contemporarily appreciatethe discrimination power (position of the patients in the plane)

www.interscience.wiley.com/journal/nbm Copyright � 200

and the functional meaning (the loadings correspond to thecorrelation coefficients of the original variables with the axes) ofthe proposed solution.From Fig. 6, it is evident that choline concentration and PE are

crucial parameters to discriminate between the benign andmalignant diseases. Choline has a high negative correlationcoefficient with LV1, while PE shows a strong negative correlationwith LV1 and LV2. Since the malignant group displays lowerscores in both LV1 and LV2 compared to benign groups, thiscorresponds to the fact that patients with adenocarcinoma havehigher values of choline and PE compared to patients withbenign pathologies. Concerning the differential diagnosis ofbenign pathologies, the descriptors endowed with the highestdiscrimination power for FB patients are the scoring of a highlevel of citrate (loadings¼ 0.460 and �0.300 on LV1 and LV2,

9 John Wiley & Sons, Ltd. NMR Biomed. 2009; 22: 1–11

119

120

T

ROO

FS

Table 1. Discriminant function analysis comparing controls versus single pathology and single pathology versus single pathologywith regard to spectroscopic 1H-MRSI or dynamic DCE-MRI or both spectroscopic and dynamic parameters

Control vs.Phlogosis

Control vs.Fibrosis

Control vs.Adenocarcinoma

Phlogosisvs. Fibrosis

Phlogosis vs.Adenocarcinoma

Fibrosis vs.Adenocarcinoma

(ChoþCre)/Cit ratioF-value 34.45 1.34 7.03 46.82 5.54 8.49Specificity 100 27 46 100 31 46Sensitivity 100 77 100 93 100 100Accuracy 100 54 71 96 67 73

Onset time, time to peak and peak enhancement parametersF-value 2.31 8.003 53.81 9.22 31.43 49.53Specificity 64 100 100 85 92 100Sensitivity 79 85 100 93 93 100Accuracy 72 92 100 89 93 100

(ChoþCre)/Cit ratio, onset time, time to peak and peak enhancement parametersF-value 12.19 6.05 39.12 15.39 22.90 38.24Specificity 100 100 100 100 92 100Sensitivity 93 85 100 93 93 100Accuracy 96 92 100 96 93 100

For each discriminant function the specificity, sensitivity and accuracy are reported.

METABOLOMIC ANALYSIS OF PROSTATIC DISEASES 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

respectively) and the presence of a comparatively high value oftime-to-peak (loadings¼ 0.308 and �0.147 on LV1 and LV2,respectively). It is worth noting that some patients with fibrosisare characterized by high values of onset time (loadings¼ 0.630and 0.133 on LV1 and LV2, respectively). Furthermore, phlogosisprovokes an increase of creatine level (high positive correlationwith LV2, loading¼ 0.605) with respect to other benignpathologies.

C

94

95

96

97

98

99

100

RRE

DISCUSSION

Due to the increased use of both serum PSA screening andTRUS-guided biopsies, prostate cancer is being identified at anearlier and more treatable stage (48). Therefore there is anincreased interest in routine check-ups, but clinical parameters

UNCO

Table 2. t-test comparing pathologic patients on the whole versusversus single pathology

Patient Group PC1 (35.43

Pathologic patients on the whole vs. controls 0.004�

Controls vs. phlogosis 0.043�

Controls vs. fibrosis 0.144Controls vs. adenocarcinoma 0.0001�

Phlogosis vs. fibrosis 0.003�

Phlogosis vs. adenocarcinoma 0.0001�

Fibrosis vs. adenocarcinoma 0.0001�

�Threshold p< 0.05.In parentheses the percent of variance explained by each principal co

NMR Biomed. 2009; 22: 1–11 Copyright � 2009 John Wiley & S

ED Palone are not sufficient to predict the course of a benign disease.In fact, the risk of over-detection has been estimated to varybetween 15 and 84% (49,50). Current classification systems areable to predict only a binary outcome, i.e. benign or malignant,with sensitivities of 95 and 73% as well as specificities of 91 and81% for 1H-MRSI and DCE-MRI, respectively (17,51). The mostused metabolic classification method is based on previouslyreported differences between cancer and normal prostate tissue;voxels are considered suspicious for cancer if the [ChoþCr]/Citratio is at least 2 SDs above the average ratio for the normalperipheral zone, and voxels are considered very suspicious forcancer if [ChoþCr]/Cit ratio is more than 3 SDs above the averageratio (21,22). However, other conditions such as prostatitis or postbiopsy haemorrhage might increase the [Choþ Cr]/Cit ratio, andalso normal prostatic tissue may show higher [ChoþCr]/Cit ratioswithin the transitional and periurethral tissue (52). Furthermore,

controls, controls versus single pathology and single pathology

Statistical Significance

) PC2 (21.73) PC3 (15.85) PC4 (14.33)

0.514 0.0001� 0.7730.207 0.004� 0.0540.050� 0.014� 0.3780.295 0.001� 0.7020.009� 0.626 0.0001�

0.039� 0.421 0.0001�

0.267 0.859 0.397

mponent is reported.

ons, Ltd. www.interscience.wiley.com/journal/nbm

7

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

T

ROO

FS

Figure 4. PCA score plot of the first two PCs (PC1 and PC2) versus thefourth PC (PC4) of the 45 patient data set. Differentiation among path-

ologies is shown.

Figure 2. PCA score plot of the third PC (PC3) versus the first PC (PC1) of

the 45 patient data set. Differentiation between control and patient (as a

whole) groups and between control and single-pathologic groups (con-trol vs. phlogosis and control vs. fibrosis) is shown.

M. VALERIO ET AL.

8

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

RRECalthough average values and SDs for [ChoþCr]/Cit in healthyprostate tissue have been available, to date, it remains unclear ifthese values are independent of the respective MR scanner and/or MRS sequence applied. In the Shukla-Dave et al. (53) study, Chowas elevated in 9 out of 12 patients with histopathologicallyconfirmed chronic prostatitis, and 86% of the voxels indicatedintermediate or high-grade diseases.The diagnostic value of DCE-MRI in histologically proven

benign and malignant prostate tissues has been evaluated byseveral studies (54,55), which postulated that prostate cancershowed earlier and stronger enhancement than normal tissue. Inparticular, Ren et al. (39) demonstrated the potential of DCE-MRIto distinguish between BPH nodules and PrC foci; the time-to-peak of PrC lesion occurred earlier than the BPH peak time andthe enhancement degree and rate of PrC were higher than thoseof BPH. On the other hand, limitations of the technique includinginadequate lesion characterization, particularly in the differen-tiation of prostatitis from cancer in the peripheral gland and in

UNCO

Figure 3. PCA score plot of the third PC (PC3) vs. the second PC (PC2) ofthe 45 patient data set. Differentiation between control and fibrosis

groups is shown.

www.interscience.wiley.com/journal/nbm Copyright � 200

ED Pthe discrimination between BPH and central gland tumours hasbeen established (34).The combined use of 1H-MRSI and DCE-MRI techniques could

be able to address the limitations found in the two techniqueswhen used independently, improving the prediction accuracy(19). van Dorsten et al. (56) showed that the addition of 1H-MRSIand DCE-MRI to the conventional MRI protocol has greatpotential for improved localization and characterization ofprostate cancer in a clinical setting.Our results represent a further improvement along this line.

Multivariate analysis is muchmore efficient in discrimination thanin the use of original variables as they do not allow for an equallyprecise discrimination: the ([Choþ Cr]/Cit) ratio, the mostdiscriminant 1H-MRSI index, fails to separate between controland FB groups, whereas dynamic parameters are not able toseparate the PL/CO groups. Furthermore, the classical, combined

Table 3. PLS-DA modelQ1 summary for discriminating1H-MRSI/DCE-MRI data from patients with both benign andmalignant prostatic diseases

Model LV R2X R2Y Q2

45 units LV1 0.353 0.273 0.249LV2 0.187 0.172 0.164LV3 0.158 0.099 0.078

41 units LV1 0.364 0.283 0.274LV2 0.155 0.185 0.159LV3 0.151 0.102 0.069

R2X, cumulative fraction of the variation of the X variableexplained per component; R2Y, cumulative fraction of thevariation of the Y variable explained per component; Q2, thecumulative predicted fraction (cross-validation) of the vari-ation joint X and Y.

Q1

9 John Wiley & Sons, Ltd. NMR Biomed. 2009; 22: 1–11

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

T

ROO

FSFigure 6. PLS-DA score and loading plots (superimposed) of the 41patient data set containing no outliers. The score plot provides a map

of how the groups relate to each other showing differentiation among

classes, while the loadings plot reveals which original variables (Cho, Cr,

Cit, OT, TP and PE) are important in separating the four groups. This figureis available in colour online at www.interscience.wiley.com/journal/nbm

Figure 5. PLS-DA score plot of the first two LVs (LV1 vs. LV2) of the 45patient data set. Differentiation among control and pathology groups is

shown.

METABOLOMIC ANALYSIS OF PROSTATIC DISEASES 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

ORREC

1H-MRSI and DCE-MRI analysis showed a lower specificity,sensitivity and accuracy compared to those obtained for thePLS model.The first multivariate data analysis applied was PCA. The

application of this method embodies a sort of ‘natural normal-ization’ of the studied data set, given that principal componentscorrespond to the eigenvectors of the correlation matrix that inturn corresponds by definition to the covariance matrix of thestandardized variables. This is particularly convenient whendealing with heterogeneous variables defined by completelydifferent measurement units (57), ruling out all questionablea priori defined standardization processes.PCA showed a natural trend to clustering of the prostatic

diseases in the MR space. This is a proof-of-concept of thepossibility of obtaining metabolic and morphological fingerprintsuseful for the differential diagnosis of prostatic diseases, evenwithout specifically imposing the discrimination task to themodel. Principal components are orthogonally constructed thusthe different relevance of the components for the discriminationof diverse prostatic diseases is the image in light of the biologicaldifferences between pathologies (58).

UNC

Table 4. Membership score (correlation coefficient) for eachof the six samples derived from the PLS model built using thewhole data set (45 units/6 variables)

Sample Control Phlogosis Fibrosis Adenocarcinoma

PL 0.233 0.428 0.126 0.212PL 0.340 0.501 0.041 0.118FB 0.323 �0.008 0.852 �0.168AC 0.048 0.107 �0.028 0.873AC 0.011 0.355 �0.058 0.693FB 0.175 0.069 0.813 �0.057

The sample column reports the effective clinical status of thepatient, the other columns the allocation coefficient made bythe system on the sole basis of MR information.

NMR Biomed. 2009; 22: 1–11 Copyright � 2009 John Wiley & S

9

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

ED PThis clustering tendency of the prostatic diseases in the MRspace was confirmed in terms of diagnostic accuracy, by PLSmethodology that highlighted choline, creatine and citrate as themain discriminant metabolites among different prostate dis-eases. The dynamic parameters endowedwith the highest clinicalsignificance were onset time, time-to-peak and PE. Takenaltogether, metabolic and dynamic descriptors allowed us toobtain a correct reclassification of an independent test set inaddition to a complete classification of the training set. Theprediction of the six patients in the test set confirms the exactdiscrimination already found in the general (45 patients) data set,adding the dimension of the generalization ability to the pureinternal consistency of the model.To investigate the metabolic and dynamic biomarkers, we

obtained a PLS model using a reduced data set by removing theoutlier (41 patients). The reproduction of the same correlationstructure, by means of a data set depurated by the most extremestatistical units, is proof of the fact that we can safely rule out a‘range restriction effect’ (46) as a possible source of confoundingfor our results. The fact that both the PLS analyses gave rise to thesame result is an important proof of the robustness of theclassification.As we mentioned in the Results section, the two latent

variables (LV1 and LV2) endowed with the highest discriminationpower were mainly related to the opposition between cholineand onset time for LV1 (these two variables are at the oppositepoles of the LV1 axis in Fig. 6) and PE and creatine for LV2(extreme opposite poles of the LV2 axis in Fig. 6). As for the LV1axis, the high choline pole matches to adenocarcinoma patients(black squares in Fig. 6), while the opposite LV pole correspondingto high values of onset time is the preferred location of fibrosispatients (black circles in Fig. 6). As for the LV2 axis, near thecreatine (Cr) pole in the LV space we observe the phlogosispatients (black triangles in Fig. 6), whereas the adenocarcinomapatients shifted toward the direction of the high PE pole. Theincreased efficiency of both PCA and PLS-DA analyses is a naturalconsequence of the fact that both techniques filter out the noisy

ons, Ltd. www.interscience.wiley.com/journal/nbm

119

120

M. VALERIO ET AL.

10

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

portion of information into minor components while concentrat-ing on the most informative portion of information retained bythe major axes (40,45).Our data are consistent with the well-known increased degree

of vascularization in tumour pathologies that scale with PE, whilethe discriminant ability of choline can be related to the changesin cell membrane synthesis and degradation of tumour tissuesthat go hand in hand with an increased choline concentration(19–26,39).Beside the mechanistic interpretation of the results, we can

safely affirm that the non-invasive acquisition of 1H-MRSI/DCE-MRI data is a potentially valid approach in both thedifferential diagnosis and treatment evaluation of prostaticdiseases.Our study can be considered as a pilot study: the paucity of the

considered data set together with the limitation of ‘pure’ diseases(e.g. patients with mixed syndromes where BPH nodules go handin hand with cancer are excluded by the analysis) are strongcaveats to the generalization of our findings. A particularly hardconstraint is the limitation to pure diseases given the highprevalence of mixed syndromes in nature. Nevertheless, in orderto validate the method, at first, we preferred to rely upon a moreneat case. Further developments of metabolomic research in thefuture will need to address this very important point (59).

T

Q2

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

UNCORREC

REFERENCES

1. Cooperberg MR, Lubeck DP, Grossfeld GD, Mehta SS, Carroll PR.Contemporary trends in imaging test utilization for prostate cancerstaging: data from the cancer of the prostate strategic urologicresearch endeavour. J. Urol. 2002; 168: 491–495.

2. Mueller-Lisse UG, Scherr MK. Proton MR spectroscopy of the prostate.Eur. Radiol. 2007; 63: 351–360.

3. Westphalen AC, McKenna DA, Kurhanewicz J, Coakley FV. Role ofmagnetic resonance imaging and magnetic resonance spectroscopicimaging before and after radiotherapy for prostate cancer.J. Endourol. 2008; 22: 789–794.

4. StephanQ2 C, Buker N, Cammann H, Meyer HA, Lein M, Jung K.Artificial neural network (ANN) velocity better identifies benign pro-static hyperplasia but not prostate cancer compared with PSAvelocity. BMC Urol. 2008; 2: 8–10.

5. Shukla-Dave A, Hricak H, Eberhardt SC, Olgac S, Muruganandham M,Scardino PT, Reuter VE, Koutcher JA, Zakian KL. Chronic prostatitis: MRimaging and 1H MR spectroscopic imaging findings-initial obser-vations. Radiology 2004; 231: 717–724.

6. Oesterling JE, Jacobsen SJ, Chute CG. Serum prostate-specific antigenin a community-based population of healthy men: establishment ofage-specific reference ranges. JAMA 1993; 270: 860–866.

7. Flanigan RC, Catalona WJ, Richie JP, Ahmann FR, Hudson ML, ScardinoPT, deKernion JB, Ratliff TL, Kavoussi LR, Dalkin BL, Waters WB,MacFarlane MT, Southwick PC. Accuracy of digital rectal examinationand transrectal ultrasonography in localizing prostate cancer. J. Urol.1994; 152: 1506–1509.

8. Eskew LA, Bare RL, McCullough DL. Systematic 5 region prostatebiopsy is superior to sextant method for diagnosing carcinoma of theprostate. J. Urol. 1997; 157: 199–203.

9. Chon CH, Lai FC, McNeal JE, Presti JC, Jr. Use of extended systematicsampling in patients with a prior negative prostate needle biopsy.J. Urol. 2002; 167: 2457–2460.

10. Pepe P, Aragona F. Prostate needle biopsy: 12 vs. 18 cores: is itnecessary? Urol. Int. 2005; 74: 19–22.

11. Samir S, Taneja MD. Prostate biopsy: targeting cancer for detectionand therapy. Rev. Urol. 2006; 8: 173–182.

12. Muller-Lisse UL, Hofstetter A. Urological diagnosis of prostate carci-noma. Radiologe 2003; 43: 432–440.

13. Okamura T, Umemoto Y, Yamashita K, Suzuki S, Shirai T, Hashimoto Y,Kohri K. Pitfalls with MRI evaluation of prostate cancer detection:

www.interscience.wiley.com/journal/nbm Copyright � 200

ED PRO

OFS

comparison of findings with histopathological assessment of retro-pubic radical prostatectomy specimens. Urol. Int. 2006; 77: 301–306.

14. Hom JJ, Coakley FV, Simko JP, Qayyum A, Lu Y, Schmitt L, Carroll PR,Kurhanewicz J. Prostate cancer: endorectal MR imaging and MRspectroscopic imaging–distinction of true-positive results from chan-ce-detected lesions. Radiology 2006; 238: 192–199.

15. Chen M, Dang HD, Wang JY, Zhou C, Li SY, Wang WC, Zhao WF, YangZH, Zhong CY, Li GZ. Prostate cancer detection: comparison ofT2-weighted imaging, diffusion-weighted imaging, proton magneticresonance spectroscopic imaging, and the three techniques com-bined. Acta Radiol. 2008; 49: 602–610.

16. Coakley FV, Hricak H, Wefer AE, Speight JL, Kurhanewicz J, Roach M.Brachytherapy for prostate cancer: endorectal MR imaging of localtreatment-related changes. Radiology 2001; 219: 817–821.

17. Scheidler J, Hricak H, Vigneron DB, Yu KK, Sokolov DL, Huang LR,Zaloudek CJ, Nelson SJ, Carroll PR, Kurhanewicz J. Prostate cancer:localization with three-dimensional proton MR spectroscopic imagi-ng-clinicopathologic study. Radiology 1999; 213: 473–480.

18. Kurhanewicz J, Swanson MG, Nelson SJ, Vigneron DB. Combinedmagnetic resonance imaging and spectroscopic imaging approach tomolecular imaging of prostate cancer. J. Magn. Reson. Imaging 2002;16: 451–463.

19. Futterer JJ, Heijmink SW, Scheenen TW, Veltman J, Huisman HJ, Vos P,Hulsbergen-Van de Kaa CA, Witjes JA, Krabbe PF, Heerschap A,Barentsz JO. Prostate cancer localization with dynamic contras-t-enhanced MR imaging and proton MR spectroscopic imaging.Radiology 2006; 241: 449–458.

20. Jung JA, Coakley FV, Vigneron DB, Swanson MG, Qayyum A, WeinbergV, Jones KD, Carroll PR, Kurhanewicz J. Prostate depiction at endor-ectal MR spectroscopic imaging: investigation of a standardizedevaluation system. Radiology 2004; 233: 701–708.

21. Males RG, Vigneron DB, Star-Lack J, Falbo SC, Nelson SJ, Hricak H,Kurhanewicz J. Clinical application of BASING and spectral/spatialwater and lipid suppression pulses for prostate cancer staging andlocalization by in vivo 3D 1H magnetic resonance spectroscopicimaging. Magn. Reson. Med. 2000; 43: 17–22.

22. Kurhanewicz J, Vigneron DB, Hricak H, Narayan P, Carroll P, Nelson SJ.Three-dimensional H-1 MR spectroscopic imaging of the in situhuman prostate with high (0.24–0.7-cm3) spatial resolution. Radi-ology 1996; 198: 795–805.

23. Heijmink SW, Futterer JJ, Hambrock T, Takahashi S, Scheenen TW,Huisman HJ, Hulsbergen-Van de Kaa CA, Knipscheer BC, Kiemeney LA,Witjes JA, Barentsz JO. Prostate cancer: body-array versus endorectalcoil MR imaging at 3 T–comparison of image quality, localization, andstaging performance. Radiology 2007; 244: 184–195.

24. Kurhanewicz J, Vigneron DB, Hricak H, Parivar F, Nelson SJ, ShinoharaK, Carroll PR. Prostate cancer: metabolic response to cryosurgery asdetected with 3D H-1 MR spectroscopic imaging. Radiology 1996;200: 489–496.

25. Costello LC, Franklin RB, Narayan P. Citrate in the diagnosis of prostatecancer. Prostate 1999; 38: 237–245.

26. Kurhanewicz J, Dahiya R, Macdonald JM, Chang LH, James TL, NarayanP. Citrate alterations in primary and metastatic human prostaticadenocarcinomas: 1H magnetic resonance spectroscopy and bio-chemical study. Magn. Reson. Med. 1993; 29: 149–157.

27. Li SY, Chen M, Wang R, Zhou C. Differentiation between benignprostatic hyperplasia and prostate cancer in the transitional zoneevaluated by 1H magnetic resonance spectroscopic imaging. Chin.Med. Sci. J. 2007; 22: 238–242.

28. Zakian KL, Sircar K, Hricak H, Chen HN, Shukla-Dave A, Eberhardt S,Muruganandham M, Ebora L, Kattan MW, Reuter VE, Scardino PT,Koutcher JA. Correlation of proton MR spectroscopic imaging withgleason score based on stepsection pathologic analysis after radicalprostatectomy. Radiology 2005; 234: 804–814.

29. McNeal JE. Normal anatomy of the prostate and changes in benignprostatic hypertrophy and carcinoma. Semin. Ultrasound CTMR 1988;9: 329–334.

30. Coakley FV, Kurhanewicz J, Lu Y, Jones KD, Swanson MG, Chang SD,Carroll PR, Hricak H. Prostate cancer tumor volume: measurementwith endorectal MR and MR spectroscopic imaging. Radiology 2002;223: 91–97.

31. Futterer JJ, Heijmink SW, Scheenen TW, Jager GJ, Hulsbergen-Van deKaa CA, Witjes JA, Barentsz JO. Prostate cancer: local staging at 3-Tendorectal MR imaging—early experience. Radiology 2005; 238:184–191.

9 John Wiley & Sons, Ltd. NMR Biomed. 2009; 22: 1–11

119

120

T

Q3

Q4

METABOLOMIC ANALYSIS OF PROSTATIC DISEASES 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

REC

32. Scheenen TW, Heijmink SW, Roell SA, Hulsbergen-Van de Kaa CA,Knipscheer BC, Witjes JA, Barentsz JO, Heerschap A. Three-dimensional proton MR Spectroscopy of human prostate at 3 Twithout endorectal coil: feasibility. Radiology 2007; 245: 507–516.

33. Carlani M, Mancino S, Bonanno E, Finazzi Agro E, Simonetti G.Combined morphological, [1H]-MR spectroscopic and contras-t-enhanced imaging of human prostate cancer with a 3-Tesla scanner:preliminary experience. Radiol. Med. (Torino) 2008; 113: 670–688.

34. Alonzi R, Padhani AR, Allen C. Dynamic contrast enhanced MRI inprostate cancer. Eur. J. Radiol. 2007; 63: 335–350.

35. Noworolski SM, Henry RG, Vigneron DB, Kurhanewicz J. Dynamiccontrast-enhanced MRI in normal and abnormal prostate tissues asdefined by biopsy, MRI, and 3D MRSI. Magn. Reson. Med. 2005; 53:249–255.

36. Padhani AR, Husband JE. Dynamic contrast-enhanced MRI studies inoncology with an emphasis on quantification, validation and humanstudies. Clin. Radiol. 2001; 56: 607–620.

37. Rouviere O, Raudrant A, Ecochard R, Colin-Pangaud C, Pasquiou C,Bouvier R, Marechal JM, Lyonnet D. Characterization of time-enhancement curves of benign and malignant prostate tissue atdynamic MR imaging. Eur Radiol. 2003; 13: 931–942.

38. Muramoto S, Uematsu H, Kimura H, Ishimori Y, Sadato N, Oyama N,Matsuda T, Kawamura Y, Yonekura Y, Okada K, Itoh H. Differentiation ofprostate cancer from benign prostate hypertrophy using dual-echodynamic contrast MR imaging. Eur. J. Radiol. 2002; 44: 52–58.

39. Ren J, Huan Y, Wang H, Chang YJ, Zhao HT, Ge YL, Liu Y, Yang Y.Dynamic contrast-enhanced MRI of benign prostatic hyperplasia andprostatic carcinoma: correlation with angiogenesis. Clin. Radiol. 2008;63: 153–159.

40. Benigni R, Giuliani A. Quantitative modeling and biology: the multi-variate approach. Am. J. Physiol. 1994; 266: R1697–R1704.

41. Franke R. Theoretical Drug Design Methods. Elsevier: Amsterdam, 1984.42. Scheenen TW, Klomp DW, Roll SA, Futterer JJ, Barentsz JO, Heerschap

A. Fast acquisition-weighted three-dimensional proton MR spectro-scopic imaging of the human prostate. Magn. Reson. Med. 2004; 52:80–88.

43. HochQ3 J, Stern A. NMR Data Processing. Wiley and Sons: 1996.44. Bartella L, Thakur SB, Morris EA, Dershaw DD, Huang W, Chough E,

Cruz MC, Liberman L. Enhancing nonmass lesions in breast: evalu-ation with proton (1H) MR spectroscopy. Radiology 2007; 245: 80–87.

45. Eriksson L, Johansson E, Kettaneh-Wold N, Wold S. Multi- and mega-variate data analysis – principles and applications. In Multi- andMegavariate Data Analysis – Principles and Applications. Eriksson L,Johansson E, Kettaneh-Wold N, Wold S (eds). Umetrics Academy,Umea: Sweden, 2001.

46. Guilford J. Fundamental Statistics in Psychology and Education.McGraw Hill: New York, 1965.

UNCOR

NMR Biomed. 2009; 22: 1–11 Copyright � 2009 John Wiley & S

ED PRO

OFS

47. LebartQ4 L, Morineau A, Warwick KM.Multivariate Descriptive StatisticalAnalysis. Wiley and Sons: 1984.

48. Wang L, Hricak H, Kattan MW, Chen HN, Kuroiwa K, Eisenberg HF,Scardino PT. Prediction of seminal vesicle invasion in prostate cancer:incremental value of adding endorectal MR imaging to the Kattannomogram. Radiology 2007; 242: 182–188.

49. Etzioni R, Penson DF, Legler JM, Di Tommaso D, Boer R, Gann PH, FeuerEJ. Overdiagnosis due to prostate-specific antigen screening: lessonsfrom U.S. prostate cancer incidence trends. J. Natl Cancer Inst. 2002;94: 981–990.

50. Draisma G, Boer R, Otto SJ, van der Cruijsen IW, Damhuis RA, SchroderFH, de Koning HJ. Lead times and overdetection due to prostate-specific antigen screening: estimates from the European RandomizedStudy of Screening for Prostate Cancer. J. Natl Cancer Inst. 2003; 95:868–878.

51. Jager GJ, Ruijter ET, van de Kaa CA, de la Rosette JJ, Oosterhof GO,Thornbury JR, Ruijs SH, Barentsz JO. Dynamic TurboFLASH subtractiontechnique for contrast-enhanced MR imaging of the prostate: corre-lation with histopathologic results. Radiology 1997; 203: 645–652.

52. Kurhanewicz J, Vigneron DB, Males RG, Swanson MG, Yu KK, Hricak H.The prostate: MR imaging and spectroscopy. Present and future.Radiol. Clin. North Am. 2000; 38: 115–138.

53. Shukla-Dave A, Hricak H, Kattan MW, Pucar D, Kuroiwa K, Chen HN,Spector J, Koutcher JA, Zakian KL, Scardino PT. The utility of magneticresonance imaging and spectroscopy for predicting insignificantprostate cancer: an initial analysis. BJU Int. 2007; 99: 786–793.

54. Padhani AR, Gapinski CJ, Macvicar DA, Parker GJ, Suckling J, Revell PB,Leach MO, Dearnaley DP, Husband JE. Dynamic contrast enhancedMRI of prostate cancer: correlation with morphology and tumourstage, histological grade and PSA. Clin. Radiol. 2000; 55: 99–109.

55. Checkley D, Tessier JJ, Kendrew J, Waterton JC, Wedge SR. Use ofdynamic contrast-enhanced MRI to evaluate acute treatment withZD6474, a VEGF signalling inhibitor, in PC-3 prostate tumours. Br. J.Cancer 2003; 89: 1889–1895.

56. van Dorsten FA, van der Graaf M, Engelbrecht MR, van Leenders GJ,Verhofstad A, Rijpkema M, de la Rosette JJ, Barentsz JO, Heerschap A.Combined quantitative dynamic contrast-enhanced MR imaging and(1)H MR spectroscopic imaging of human prostate cancer. J. Magn.Reson. Imaging 2004; 20: 279–287.

57. Sirabella P, Giuliani A, Colosimo A, Dippner JW. Deconvolving theclimate effects on cod recruitment by principal component analysisand canonical correlation. Mar. Ecol. Prog. Ser. 2001; 218: 213–222.

58. Crescenzi M, Giuliani A. The main biological determinants of tumorline taxonomy elucidated by a principal component analysis ofmicroarray data. FEBS Lett. 2001; 507: 114–118.

59. Nicholson JK, Lindon JC. Systems biology: metabonomics. Nature2008; 455: 1054–1056.

98

ons, Ltd. www.interscience.wiley.com/journal/nbm

11

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

M. Valerio, V. Panebianco, A.Sciarra, M. Osimani, S. Salsiccia,L. Casciani, A. Giuliani, M. Bizzarri,F. Di Silverio, R. Passariello, F.Conti* . . . . . . . . . . . . . . . . . xx–xx

ClassificationQ5 of prostatic diseasesby means of multivariate analysis on invivo proton MRSI and DCE-MRI data

Q5

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

UNCORRECTE

D PRO

OFS

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

Author Query Form (NBM/1408) 64

65

66

67

Special Instruction: Author please include responses to queries with your other corrections and

return by e-mail.

68

69

70

71

72

73

74

75

76

77

78

79

FS

Q1: Author: Please check the presentation of the table.

Q2: Author: Reference is not cited in the text, please check.

Q3: Author: Please provide publisher location.

Q4: Author: Please provide publisher location.

Q5: Author: Please supply a graphical abstract (GA). This consists of a 70-word abstract plusfigure. The figure can be one of those already in the paper or another of your own choosing. TheGA accompanies the online version of the paper and appears in the table of contents when thepaper is assigned to an issue and printed.

O 80

UNCORRECTE

D PRO 81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

WILEY AUTHOR DISCOUNT CLUB We would like to show our appreciation to you, a highly valued contributor to Wiley’s publications, by offering a unique 25% discount off the published price of any of our books*. All you need to do is apply for the Wiley Author Discount Card by completing the attached form and returning it to us at the following address:

The Database Group (Author Club) John Wiley & Sons Ltd The Atrium Southern Gate Chichester PO19 8SQ UK

Alternatively, you can register online at www.wileyeurope.com/go/authordiscount Please pass on details of this offer to any co-authors or fellow contributors. After registering you will receive your Wiley Author Discount Card with a special promotion code, which you will need to quote whenever you order books direct from us. The quickest way to order your books from us is via our European website at:

http://www.wileyeurope.com Key benefits to using the site and ordering online include: • Real-time SECURE on-line ordering • Easy catalogue browsing • Dedicated Author resource centre • Opportunity to sign up for subject-orientated e-mail alerts Alternatively, you can order direct through Customer Services at: cs-books@wiley.co.uk, or call +44 (0)1243 843294, fax +44 (0)1243 843303 So take advantage of this great offer and return your completed form today. Yours sincerely,

Verity Leaver Group Marketing Manager author@wiley.co.uk

*TERMS AND CONDITIONS This offer is exclusive to Wiley Authors, Editors, Contributors and Editorial Board Members in acquiring books for their personal use. There must be no resale through any channel. The offer is subject to stock availability and cannot be applied retrospectively. This entitlement cannot be used in conjunction with any other special offer. Wiley reserves the right to amend the terms of the offer at any time.

To enjoy your 25% discount, tell us your areas of interest and you will receive relevant catalogues or leaflets from which to select your books. Please indicate your specific subject areas below.

Accounting • Public • Corporate

[ ] [ ] [ ]

Architecture Business/Management

[ ] [ ]

Chemistry • Analytical • Industrial/Safety • Organic • Inorganic • Polymer • Spectroscopy

[ ] [ ] [ ] [ ] [ ] [ ] [ ]

Computer Science • Database/Data Warehouse • Internet Business • Networking • Programming/Software

Development • Object Technology

[ ] [ ] [ ] [ ] [ ] [ ]

Encyclopedia/Reference • Business/Finance • Life Sciences • Medical Sciences • Physical Sciences • Technology

[ ] [ ] [ ] [ ] [ ] [ ]

Engineering • Civil • Communications Technology • Electronic • Environmental • Industrial • Mechanical

[ ] [ ] [ ] [ ] [ ] [ ] [ ]

Earth & Environmental Science Hospitality

[ ] [ ]

Finance/Investing • Economics • Institutional • Personal Finance

[ ] [ ] [ ] [ ]

Genetics • Bioinformatics/ Computational Biology • Proteomics • Genomics • Gene Mapping • Clinical Genetics

[ ] [ ] [ ] [ ] [ ] [ ]

Life Science Landscape Architecture Mathematics Statistics Manufacturing Materials Science

[ ] [ ] [ ] [ ] [ ] [ ]

Medical Science • Cardiovascular • Diabetes • Endocrinology • Imaging • Obstetrics/Gynaecology • Oncology • Pharmacology • Psychiatry

[ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ]

Psychology • Clinical • Forensic • Social & Personality • Health & Sport • Cognitive • Organizational • Developmental & Special Ed • Child Welfare • Self-Help

[ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ] [ ]

Non-Profit [ ] Physics/Physical Science [ ]

Please complete the next page /

REGISTRATION FORM For Wiley Author Club Discount Card

I confirm that I am (*delete where not applicable): a Wiley Book Author/Editor/Contributor* of the following book(s): ISBN: ISBN: a Wiley Journal Editor/Contributor/Editorial Board Member* of the following journal(s): SIGNATURE: …………………………………………………………………………………… Date: ………………………………………

PLEASE COMPLETE THE FOLLOWING DETAILS IN BLOCK CAPITALS: TITLE: (e.g. Mr, Mrs, Dr) …………………… FULL NAME: …………………………………………………………………………….… JOB TITLE (or Occupation): ..………………………………………………………………………………………………………………… DEPARTMENT: …………………………………………………………………………………………………………………………………………….. COMPANY/INSTITUTION: …………………………………………………………………………………………………………………………… ADDRESS: …………………………………………………………………………………………………………………………………………………… ……………………………………………………………………………………………………………………………………………………………………… TOWN/CITY: ………………………………………………………………………………………………………………………………………………… COUNTY/STATE: …………………………………………………………………………………………………………………………………………. COUNTRY: ……………………………………………………………………………………………………………………………………………………. POSTCODE/ZIP CODE: ………………………………………………………………………………………………………………………………… DAYTIME TEL: ……………………………………………………………………………………………………………………………………………… FAX: ……………………………………………………………………………………………………………………………………………………………… E-MAIL: …………………………………………………………………………………………………………………………………………………………

YOUR PERSONAL DATA We, John Wiley & Sons Ltd, will use the information you have provided to fulfil your request. In addition, we would like to:

1. Use your information to keep you informed by post of titles and offers of interest to you and available from us or other Wiley Group companies worldwide, and may supply your details to members of the Wiley Group for this purpose.

[ ] Please tick the box if you do NOT wish to receive this information

2. Share your information with other carefully selected companies so that they may contact you by post with details of titles and offers that may be of interest to you.

[ ] Please tick the box if you do NOT wish to receive this information. E-MAIL ALERTING SERVICE We also offer an alerting service to our author base via e-mail, with regular special offers and competitions. If you DO wish to receive these, please opt in by ticking the box [ ].

If, at any time, you wish to stop receiving information, please contact the Database Group (databasegroup@wiley.co.uk) at John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, PO19 8SQ, UK.

TERMS & CONDITIONS This offer is exclusive to Wiley Authors, Editors, Contributors and Editorial Board Members in acquiring books for their personal use. There should be no resale through any channel. The offer is subject to stock availability and may not be applied retrospectively. This entitlement cannot be used in conjunction with any other special offer. Wiley reserves the right to vary the terms of the offer at any time. PLEASE RETURN THIS FORM TO: Database Group (Author Club), John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, PO19 8SQ, UK author@wiley.co.uk Fax: +44 (0)1243 770154