Effectiveness of salvage selective and modified radical neck dissection for regional pathologic...

ORIGINAL ARTICLE

EFFECTIVENESS OF SALVAGE SELECTIVE AND MODIFIEDRADICAL NECK DISSECTION FOR REGIONAL PATHOLOGICLYMPHADENOPATHY AFTER CHEMORADIATION

Lisa van der Putten, MD,1,2 Guido B. van den Broek, MD,1,3 Remco de Bree, MD, PhD,2

Michiel W. M. van den Brekel, MD, PhD,1,3 Alfons J. M. Balm, MD, PhD,1,3

Frank J. P. Hoebers, MD, PhD,4 Patricia Doornaert, MD,5 C. Rene Leemans, MD, PhD,2

Coen R. N. Rasch, MD, PhD4

1 Department of Head and Neck Oncology and Surgery, The Netherlands Cancer Institute–Antoni VanLeeuwenhoek Hospital, Amsterdam, The Netherlands. E-mail: [email protected] Department of Otolaryngology–Head and Neck Surgery, VU University Medical Center, Amsterdam,The Netherlands3 Department of Otolaryngology, Academic Medical Center Amsterdam, Amsterdam, The Netherlands4 Department of Radiation Oncology, The Netherlands Cancer Institute–Antoni Van Leeuwenhoek Hospital,Amsterdam, The Netherlands5 Department of Radiation Oncology, VU University Medical Center, Amsterdam, The Netherlands

Accepted 1 August 2008Published online 8 January 2009 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/hed.20987

Abstract: Background. The objective of this study was

to evaluate the effectiveness and safety of our careful obser-

vational strategy and neck dissections and the accuracy of

ultrasound-guided fine-needle aspiration cytology, and to

determine the prognostic factors for outcome and regional

control after primary chemoradiation. Diagnostic evaluation of

the regional status after concurrent chemoradiation for

advanced head and neck cancer remains difficult, and the

indications for a salvage neck dissection and its extent are not

clearly defined.

Methods. In a series of 540 patients, there was suspicion

of regional residual or recurrent disease after chemoradiation

in 61 patients who underwent 68 salvage neck dissections and

68 patients who were considered unresectable. For the

patients with salvage neck dissection, the accuracy of ultra-

sound-guided fine-needle aspiration cytology was determined.

Disease control in the neck, disease-specific and overall sur-

vival, and parameters that may have prognostic value for the

outcome were evaluated.

Results. Neck dissection specimens contained viable tu-

mor in 26 (43%) patients. Of these, 13 had selective neck dis-

sections and 13 modified radical neck dissections. Ultrasound-

guided fine-needle aspiration cytology had a sensitivity of 80%

and specificity of 42%. Nine patients developed a regional re-

currence after salvage neck dissection (5 located in contralat-

eral neck). Five-year regional control and overall survival rates

were 79% and 36%, respectively. Significant prognostic factors

for overall survival were surgical margins and ‘‘residual versus

recurrent disease’’ in multivariate analysis.

Conclusion. Considering the good regional control rate

and the high rate of unnecessary neck dissections with a

theoretical planned neck dissection strategy, we conclude

that a careful observational strategy is worthwhile and safe.

For the evaluation of radiation treatment response, ultra-

sound-guided fine-needle aspiration cytology has a low

specificity. VVC 2009 Wiley Periodicals, Inc. Head Neck 31:

593–603, 2009

Correspondence to: A. J. M. Balm

VVC 2009 Wiley Periodicals, Inc.

Neck Dissection after Chemoradiation HEAD & NECK—DOI 10.1002/hed May 2009 593

Keywords: neck dissection; salvage; chemoradiation; head and

neck cancer; advanced stage

A combination of concurrent chemotherapy andradiotherapy has become the treatment of choicefor many patients with advanced head and necksquamous cell carcinoma, with significantlyimproved locoregional control and overall sur-vival compared with radiotherapy alone.1,2

These patients were often seen with advanceddisease in the neck. Treating these patients goesalong with several dilemmas. One of theseincludes the decision whether a planned neckdissection, for pretreatment N2–3 disease in theneck, or a salvage neck dissection for residualdisease in the neck should be performed.3

Although postchemoradiation planned neck dis-section is routine in many institutions, it is of-ten considered as ‘‘overtreatment’’ because inthe majority of neck dissection specimens noviable tumor is found.4–9 Furthermore, it deliv-ers additional morbidity to patients who havealready been subjected to severe toxicity duringchemoradiation.10–12 As a consequence, there isa tendency to perform postchemoradiation neckdissections only if indicated by posttreatmentdiagnostic (clinical, radiologic, and/or cytologic)evaluation of the neck.13–17

The advantage of limiting neck dissection topatients with residual neck disease 6 to 8 weeksposttreatment is that overtreatment is reduced.However, chemoradiation interferes with a reli-able assessment of regional response because oftreatment-induced fibrosis, necrotic lymph nodeswithout tumor, and false cytologic results. Particu-larly in patients with substantial nodal tumorvolumes, differentiation between scarred and par-tially necrotic lymph nodes tissue and residualmetastases can be difficult.3,18 If the decision toperform a neck dissection has been made, the nextdilemma is determined by the extent of the neckdissection that needs to be performed, because evi-dence is emerging that selective neck dissection(SND) may be sufficient if it is highly likely that itremoves the metastatic process entirely.19,20

Over the years, ultrasound-guided fine-nee-dle aspiration cytology in combination with pal-pation and routine posttreatment CT or MRIhas been used as diagnostic tools in our insti-tutes to decide whether a salvage neck dissec-tion is indicated. For untreated necks, thereported accuracy of ultrasound-guided fine-needle aspiration cytology is high, but, to ourknowledge, no previous studies have been done

to evaluate the accuracy of ultrasound-guidedfine-needle aspiration cytology in the detectionof lymph node metastases after nonsurgicaltreatment.21–25

In this study, we investigated the effective-ness and safety of our careful observationalstrategy and neck dissection as determined byregional control and overall survival and thereliability of ultrasound-guided fine-needle aspi-ration cytology. We also determined the prognos-tic factors for outcome.

MATERIALS AND METHODS

From November 1996 until November 2005, 540patients with advanced head and neck squamouscell carcinoma were treated with concomitantchemoradiation in The Netherlands Cancer Insti-tute and the VU University Medical Center. Thisretrospective study included patients who weretreated by 5 different schemes of chemoradiation.Two hundred seven patients were treated accord-ing to the intraarterial chemoradiation scheduleconsisting of 4 consecutive weekly selective intra-arterial infusions of cisplatin (150 mg/m2), fol-lowed by intravenous sodium thiosulfate rescuecombined with simultaneous radiotherapyaccording to the RADPLAT protocol.26–28 Onehundred sixty-one patients were treated withconcomitant intravenous administration of 3 �100 mg/m2 on day 1, 22, and 43, and 119 patientswith a low-dose intravenous concomitant scheme(daily 6 mg/m2 cisplatin, 20 courses).29 Allpatients were irradiated daily for 6 to 7 weeks toa total dose of 70 Gy (2 Gy per fraction, 5–6/wk).Fifty-three patients were treated according to theEuropean Organization for Research and Treat-ment of Cancer (EORTC) 24954 trial,30 24 withan alternating scheme (cisplatin 20 mg/kg and 5-FU 200 mg/kg IV in week 1, 4, 7, 10; radiotherapyin week 2, 3, 5, 6, 8, 9, total dose 60 Gy), and 29with a sequential scheme (cisplatin 100 mg/kgand 5-FU 1000 mg/kg IV, 4 courses, followed by 7weeks radiotherapy, total dose 70 Gy). Both sidesof the neck were irradiated in all patients,regardless of the lymph node status. Four hun-dred eight patients had evidence of neck node me-tastases before treatment. Treatment evaluationwas performed 6 to 8 weeks after completion ofchemoradiation. To measure response to therapy,the treatment results were evaluated by clinicalexamination, ultrasound-guided fine-needle aspi-ration cytology, MRI, and/or CT. During routineregular follow-up, imaging or ultrasound-guided

594 Neck Dissection after Chemoradiation HEAD & NECK—DOI 10.1002/hed May 2009

fine-needle aspiration cytology was repeated ifindicated. Positive cytology was defined by atleast strong suspicion of viable tumor cells in thesmear (all other cytologic findings, eg, necrotic[tumor] cells, were defined as negative). Immuno-cytochemical staining was not used. All patientswith suspicion of persistent or recurrent lymphnode metastases and who were considered opera-ble underwent a neck dissection. Lymph node me-tastases were defined persistent when diagnosedwithin 3 months and recurrent when diagnosedat least 3 months after chemoradiation. Medianfollow-up was 18 months (range, 0–98 months).Minimum follow-up after completion of the che-moradiation was 1 year or until the patient died.Neck dissections were performed 6 to 87 weeksafter completion of chemoradiation.

The extent of the neck dissection was basedon posttreatment clinical and radiologic evalua-tion. SND (levels I–III or II–IV) was performedfor removal of limited residual mass in the neck;a modified radical neck dissection (MRND) wasperformed for removal of multiple or extensiveresidual metastases.

Using follow-up and histopathologic examina-tion of the neck dissection specimen as referencestandard, the sensitivity, specificity, positive andnegative predictive values, and overall accuracywere calculated for ultrasound-guided fine-needleaspiration cytology. Two independent groupswere compared using the chi-square test. Re-gional control and overall survival were calcu-lated by the Kaplan–Meier method from the dayof the neck dissection. Univariate analysis (log-rank test) was performed to determine the pre-dictive value of the following variables forregional control and overall survival: primary tu-mor site, pretreatment T classification, pretreat-ment N classification, chemoradiation schedule,extent of neck dissection, pathologic examination

of the surgical specimen, and residual versusrecurrent disease (�3 or >3 months after chemo-radiation). A multivariate analysis (Cox regres-sion, Wald test) was performed to determineindependent factors for overall survival after con-current chemoradiation. Statistical analyseswere carried out using SPSS 14.0.

RESULTS

Unresectable Recurrences/Residual Disease

Group. Of the 540 patients, 68 had an unresect-able regional residue (n ¼ 40) or unresectableregional recurrence (n ¼ 28) after chemoradia-tion. Of 68 patients, 18 had N3 disease. Twenty-three of 28 patients with a regional recurrencewere considered unresectable because they alsohad an unresectable local recurrence (fixation ofthe deep muscles of the neck or prevertebral fas-cia, invasion of base of skull, involvement of Eu-stachian tube, or invasion of vertebrae; n ¼ 15)or distant metastases (n ¼ 8). Three patientshad an unresectable regional recurrence(encasement of common or internal carotid ar-tery) without local or distant disease. Two otherpatients were found inoperable because of apoor general physical condition. The intervalbetween chemoradiation and presentation of theunresectable recurrence varied from less than 3months (n ¼ 3), 3 to 6 months (n ¼ 4), 6 to 12months (n ¼ 9), 12 to 24 months (n ¼ 9) to morethan 24 months (n ¼ 3).

Pretreatment Staging Versus Recurrences. Thechance of developing a regional recurrence orpersistent disease was 11% for a N0-N1 neckand 32% for a N2-N3 neck (Table 1, Figure 1).Of the patients with a N0-N1 neck before treat-ment, 7 (3.5%) eventually underwent a neck

Table 1. Classified per N classification before chemoradiation: total number of patients with chemoradiation and number and

percentage of regional residues and recurrences (total, with and without salvage neck dissection).

N classification

Total (no. of

patients)

Recurrences total

Recurrences treated by

neck dissection

Recurrences not treated

by neck dissection

No. of patients (%) No. of patients (%) No. of patients (%)

N0 132 6 (4.5) 3 (2.3) 3 (2.3)

N1 70 16 (22.9) 4 (5.7) 12 (17.1)

N2a 28 13 (46.4) 9 (32.1) 4 (14.3)

N2b 105 27 (25.7) 12 (11.4) 15 (14.3)

N2c 146 35 (24.0) 19 (13.0) 16 (11.0)

N3 59 32 (54.2) 14 (23.7) 18 (30.5)

Total 540 129 (23.9) 61 (11.3) 68 (12.6)


dissection, in comparison to 54 (16%) of thepatients with a N2-N3 neck.

Neck Dissection Group. In total, 68 neck dissec-tions for residual or recurrent disease in the neckin 61 patients were performed: 42 SNDs and 26MRNDs. The groups with SND and MRND didnot show a statistical difference in prechemora-diation N classification or diameter of the largestresected lymph node. The median intervalbetween the last radiation and the salvage neckdissection was 14 weeks (range, 6–87 weeks).Forty-five patients underwent neck dissection forsuspicion of residual and 16 patients for recur-rent disease, which was an SND in 30 patients(67%) and 7 patients (44%), respectively. Aftersalvage surgery, the median follow-up period was27 months (range, 0–95 months). Patient charac-teristics are summarized in Table 2.

Indications for neck dissections were a pal-pable mass and/or suspicion of persistent orrecurrent lymph node metastases on MRI,CT, or ultrasound-guided fine-needle aspirationcytology.

In 35 of 61 patients (41 necks), no viable tu-mor was detected at pathology. In 25 of these,necrotic remnants of metastases were found. Re-sidual or recurrent squamous cell carcinomawas histologically demonstrated in 26 patients(43%) (18 residual disease, 8 recurrent disease,p ¼ .49), ie, 27 neck dissection specimens (14SND, 13 MRND). In 23 of 26 patients (88%), themetastases were found in the same levels asbefore treatment, representing the location ofthe largest lymph node metastases in 21 ofthese patients (81%). In 3 patients (12%) themetastases were found outside the original met-astatic level as judged clinically and radiologi-cally. All patients except 1 had residual orrecurrent metastases in levels II–IV. Only 1(2%) patient had a lymph node metastasis inlevel V. Thirteen patients with SNDs (35%) and13 with MRNDs (54%) had a neck dissectionthat contained viable tumor, which was not asignificant difference (chi-square test, p ¼ .142).

At pathologic examination, negative marginswere obtained in 75% of patients with a tumorcontaining SND and 62% of patients with a pos-itive MRND (trend test, p ¼ .74). Of 61 patients

FIGURE 1. Flow diagram regional neck nodes.


who underwent salvage neck dissection, 3patients had residual disease after incompleteresection. Six others developed a regional recur-rence after neck dissection. In only 1 of these 6patients, evidence for viable tumor cells hadbeen found at histopathologic examination ofthe previous neck dissection. In the other 5patients, only necrotic tissue was found in theneck mass. In 3 patients, the regional recur-rence was the only site of recurrent tumor,whereas in the other 3, synchronously or meta-chronously a local tumor recurrence occurred.

In 4 of these 9 patients, of whom 3 had anincomplete resection, the disease persisted or

recurred in the operated neck. This concerned ahistopathologic irradical MRND in 2 patients,an irradical SND which was found unresectableintraoperatively, and a SND without tumor inthe specimen but with a recurrence more then ayear after the neck dissection. Five otherpatients developed a regional recurrence locatedin the contralateral neck, of whom 3 had origi-nally a N2c neck with contralateral lymph nodemetastases ranging from 0.9 to 1.0 cm. One ofthese 5 was treated with a contralateral neckdissection, and the other 4 were unresectable.

When reviewing the indication for an SNDor MRND, 41 heminecks (60%) were unneces-sary because no viable tumor was found in theresection specimen and no regional recurrencedeveloped in the neck (except 1 patient with aregional recurrence 1 year after neck dissec-tion). Twenty-seven heminecks contained viabletumor (Table 3). Two MRNDs had close marginswithout regional recurrence, indicating that anMRND was indeed correctly performed. Tenpatients with an SND had negative marginswithout regional disease in the follow-up, andfor these patients an extended resection of anMRND seems not necessary. In the 7 patientswith positive margins, a more extended resec-tion seems not indicated because: (a) a compre-hensive neck dissection was already performed,or (b) patients were diagnosed with a local re-currence or distant metastases within 2 months.For 8 patients with a MRND containing positivelymph nodes in selected levels and without re-gional recurrence, an SND might be indicated inretrospect.

Model of Planned Neck Dissection Strategy. Toretrospectively estimate what the effect of aplanned neck dissection strategy would havebeen, we developed a model. Patients with unre-sectable residual disease or a poor general phys-ical condition could not have undergone aplanned neck dissection and were thereforeexcluded from the calculations (42 patients).The other patients with unresectable recur-rences might have benefited from a plannedneck dissection, just as the patients with anincomplete resection during salvage neck dissec-tion (with neck dissections 4 to 9 months afterchemoradiation). This group with possible bene-fit from a planned neck dissection strategy con-stitutes 5.8% of all patients (29/498); 5.2% inthe N0-N1 group and 6.2% in the N2-N3 group(see Figure 1). The percentage of unnecessary

Table 2. Characteristics of 61 patients with salvage neck

dissection.

Variables

No. of

patients

% of

patients

Sex

Male 51 84

Female 10 16

Stage of disease

III 4 7

IV 57 93

T classification (before chemoradiation)

T1 1 2

T2 8 13

T3 22 36

T4 30 49

N classification (before chemoradiation)

N0/1 7 11

N2 9 15

N2b 12 20

N2c 19 31

N3 14 23

Primary tumor site

Oral cavity 6 10

Oropharynx 34 56

Hypopharynx 16 26

Larynx 5 8

Chemoradiation schedule

Intraarterial 27 44

Intravenous 19 31

Low dose 12 20

EORTC alternating 1 2

EORTC sequential 2 3

Residual/recurrent disease*

Residual disease 45 74

Recurrent disease 16 26

Ultrasound fine-needle aspiration cytology

Positive 30 65

Negative 16 35

Type of neck dissection

Selective neck dissection 42 62

Modified radical neck dissection 26 38

*Residual disease ¼ persistent lymph node metastases diagnosedwithin 3 months after chemoradiation; recurrent disease ¼ recurrentlymph node metastases diagnosed at least 3 months afterchemoradiation.


neck dissections (regional disease-free patients)would have been 82.5% for the total group (411/498) and 92.8% and 76% for the N0-N1 and N2-3 group, respectively. For the patients whounderwent a neck dissection (incomplete resec-tions not included) for a regional residual dis-ease or recurrence (11.7% of the total group [58/498]; 2.0% N1-N2, 17.8% N2-N3), no clear bene-fit or disadvantage was found retrospectively.

Ultrasound-Guided Fine-Needle Aspiration Cyto-

logy. Of the 61 patients, positive ultrasound-guided fine-needle aspiration cytology resultswere achieved in 30 and negative results in 16patients. In 15 patients no ultrasound-guidedfine-needle aspiration cytology was performed asthe clinical or imaging evidence of metastasiswas convincing.

Twelve (40%) of 30 patients with positiveultrasound-guided fine-needle aspiration cytol-ogy had histologically proven residual disease inthe neck dissection specimen (true positive).With the ultrasound-guided fine-needle aspira-tion cytology 3 (19%) of 16 negative aspirateswere false negative. The sensitivity, specificity,positive predictive value, negative predictivevalue, and overall accuracy of ultrasound-guidedfine-needle aspiration cytology were 80%, 42%,40%, 81%, and 57%, respectively (Table 4).There was no significant relation between thetime interval between radiotherapy and ultra-sound-guided fine-needle aspiration cytologyand the presence (mean, 15.5 weeks) or absence(mean, 14.8 weeks) of viable tumor in the neckdissection specimen (independent t test, p ¼.91).

Table 3. Retrospective indication of type of neck dissection, based on number and level of positive lymph nodes in resection

specimen, resection margins, and regional residual and recurrent disease.

Patient

SND

MRND

Lymph nodes

(number)

Lymph nodes

(level) Margins

Residue

ipsilateral

Recurrence

ipsilateral Remarks

Retrospective

SND/MRND

3 MRND 2 2, 3 Close No No ¼ MRND

4 SND 1 2, 3 Negative No No ¼ SND

5 MRND 1 Unknown Close No No ¼ MRND

12 SND 1 3 Negative No No ¼ SND

13 MRND 1 3, 4 Negative No No MRND ! SND?



23 MRND 7 1, 2, 3, 4 Positive No No Distant metastases,

1 month

Palliative*




28 SND (R) Unknown 2, 3, 4 Positive No No Distant metastases,

2 months

Palliative*

28 SND (L) Unknown Diffuse Positive No No Distant metastases,

2 months

Palliative*

34 SND 2 þ convolute 2, 3, 4 Positive No No Local recurrence, death,

6 months

Palliative*

35 SND Unknown 3 Negative No No ¼ SND

39 MRND 1 4 Negative No No MRND ! SND?

43 SND 0 Unknown Positive Yes No Intraoperatively

unresectable

Palliative*



48 MRND 1 2 Positive Yes No Palliative*

49 MRND 7 2, 3, 4 Positive Yes No Palliative*



57 MRND Fibrotic lesion 2, 3, 4 Negative No No MRND ! SND?

58 MRND 2 2, 3 Negative No No MRND ! SND?

59 MRND 3 2, 3, 4 Negative No No MRND ! SND?

61 MRND 4 2, 3 Negative No No (M)RND ! SND?

Abbreviations: R, right hemineck; L, left hemineck.*Palliative ¼ positive margins without consequences for type of neck dissection because of: (a) unresectable regional disease, or (b) death by localdisease or distant metastases within 2 months.


Survival and Regional Disease-Free Survival. Ofthe whole patient group of 540 patients, afterchemoradiation, 76% (411/540) remained free ofregional disease during follow-up (5-year re-gional control rate of 71%, analyzed from thelast radiation date). The 5-year regional controlfrom the date of salvage neck dissection for theneck dissection study group of 61 patients was79%. The 5-year regional control was 77% afterSND and 90% after MRND, but this differencewas not statistically significant (log-rank test, p¼ .70), and neither was the difference betweenresidual (77%) and recurrent disease (86%) (log-rank test, p ¼ .77). All other variables were notpredictive for regional control as well. Therefore,a multivariate analysis was not performed.

The 5-year overall survival for the totalstudy group was 36% (see Figure 2). Univariateanalysis demonstrated that the type of neck dis-section (log-rank test, p ¼ .04), histologic dem-onstration of viable squamous cell carcinoma inthe neck dissection (log-rank test, p ¼ .03; Fig-ure 3) and surgical margins (log-rank test, p <.001) were significant prognostic factors for over-all survival (Table 5). In a multivariate Cox-regression analysis with the variables: tumor vi-ability, pretreatment T and N classification, typeof neck dissection, surgical margins, and chemo-radiation scheme, the surgical margins and‘‘residual or recurrent disease’’ emerged as theonly significant prognostic factors for overallsurvival (p < .001; hazard ratio, 0.098; 95% CI,

Table 4. Diagnostic accuracy of ultrasound-guided fine-needle aspiration cytology.

Pathology neck

dissection

þ � Total

Ultrasound-guided fine-needle

aspiration cytology

þ 12 18 30

� 3 13 16

Total 15 31 46

Sensitivity Specificity PPV NPV Overall accuracy

Ultrasound-guided fine-needle

aspiration cytology

80% 42% 40% 81% 57%

Abbreviations: PPV, positive predictive value; NPV, negative predictive value.

FIGURE 2. Overall survival of selective neck dissection (SND)

group, modified radical neck dissection (MRND) group, and all

patients. Log-rank test comparing the SND group and MRND

group was statistically significant (p ¼ .04).

FIGURE 3. Overall survival of patients with viable tumor in

specimen and patients without viable tumor in specimen. Log-

rank test comparing both groups was significantly different (p ¼.03).


0.040–0.240; and p ¼ .03, respectively). Patientswith recurrent disease had better outcome thanpatients with residual disease. The presence ofviable tumor cells and the type of neck dissec-tion both showed a trend when analyzed in thesubgroup of patients with negative surgical mar-gins, but lost their significance in multivariateanalysis of the total group (p ¼ .61 and p ¼ .06,respectively).

DISCUSSION

One of the major controversies today in headand neck oncology concerns the question ofplanned neck dissection or a salvage procedureafter chemoradiation. Currently there are nowell-designed randomized trials available thatdemonstrate the benefit of 1 of the 2 surgicalapproaches. A tendency of beneficial effects ofplanned neck dissection has been observed, butthis finding is not supported by strong statisti-cal significance.31 Goguen et al32 retrospectivelyinvestigated 55 patients in a nonrandomizedway, who underwent a neck dissection after con-current chemoradiation. Patients with N2 dis-ease had no benefit of a routine planned neckdissection. When patients with and without aneck dissection after a complete clinicalresponse (indicated by physical examination andimaging studies) were compared, there was nosurvival benefit performing a planned neck dis-section. Several years ago, Mendenhall et al33

reported that planned neck dissection was notneeded for patients with N0-N1 disease whounderwent radiotherapy or for patients with N2disease, when chemotherapy was added toradiotherapy. This was confirmed by Argiris

et al.34 They found better results for plannedneck dissection only in patients with N3 diseaseor a salvage neck dissection in patients withoutclinical complete response. These studies under-line the importance to select patients for a neckdissection, but it remains unclear whether N2disease should be the cutoff point for plannedsurgery. In our study, pretreatment N classifica-tion seemed to be correlated with the chance todevelop a neck recurrence. Thirty-two percent ofthe patients with an N2-3 neck pretreatmenthad residual or recurrent regional disease, com-pared with 11% of the patients with an N0-1neck. Richey et al35 reported that patients withinitial N3 disease had a lower survival followingattempted salvage surgery. We could not confirmthis in our study.

In total, 129 of 540 patients had a neck re-currence, of whom 61 were suitable for a sal-vage neck dissection and 68 patients wereconsidered unresectable at the time of decisionfor surgery. In 42 of these 68 patients, a plannedneck dissection would not have prevented theneck disease to become unresectable because: (a)the regional residual disease was unresectableat the end of chemoradiation or (b) the patientwas inoperable because of severe comorbidity.For the other 26 patients, it is difficult to assesswhether they would have benefited from aplanned neck dissection, but in these patientsearly neck dissection could possibly have im-proved resectability. While early detection of 15local recurrences in this group of 26 patientsmight have changed the treatment for patientswith initially functionally unresectable disease,this would not have changed the treatment fortechnically unresectable disease. In 8 otherpatients, distant metastases occurred at a laterstage as well, and it is impossible to estimatewhether these could have been prevented byearly neck treatment.

When analyzed in our model, this resulted ina percentage of 6% who might have benefitedfrom a planned neck dissection, while thisplanned neck dissection would have been un-necessary in 76% of the patients with N2-N3disease. For patients with N0-N1 neck the num-ber of unnecessary neck dissections is evenhigher (92.8%). Together with the relativelygood regional control rate, this leads us to theconclusion that our watch and careful observa-tional strategy has an acceptable outcome, andthat a planned neck dissection strategy wouldhave resulted in considerable overtreatment.

Table 5. Univariate analysis and multivariate analysis for

overall survival.

Variable

p value

Univariate Multivariate

T classification .75 .81

N classification .33 .47

Chemoradiation scheme .17 .43

Tumor site .96 .29

Tumor stage .70 .70

Residual/recurrent disease .08 .03

Viable tumor .03 .61

Type of neck dissection .04 .06

Surgical margins <.001 <.001


Although in this series a relatively good re-gional control rate is obtained, we realize that,on one hand, probably still some unnecessaryneck dissections were performed and, on theother hand, in some patients a delay hasoccurred by not performing a routine neck dis-section. Especially in N2-3 disease, the chanceof a regional residue or recurrence is over 30%and such a high risk should be considered in de-cision making. As a consequence, in thesepatients we now perform a neck dissection incase of doubt on the response in the neck atimaging or clinical examination. We do not rec-ommend routine planned neck dissections incase of a complete response because this wouldresult in a high percentage of unnecessary neckdissection.

In this study, patients with neck dissectionsfor both residual and recurrent regional diseasewere included. Residual disease was shown tobe a significant independent predictor for aworse overall survival and might influence deci-sion making for planned neck dissection, realiz-ing that regional control and percentage ofviable tumor containing neck dissection speci-mens were not significantly different frompatients with recurrent disease.

It is surprising that only 8 of 16 patientswho showed evidence of recurrent disease hadhistologically positive tumor in the specimen.Clinical evidence of a suspicious neck swellingcould be explained by postchemoradiation swel-ling of lymph nodes, necroses, and false-positiveultrasound-guided fine-needle aspiration cytol-ogy. The possible severe consequences of missingand not treating a regional recurrence will alsoplay a role, by decreasing the threshold to per-form a neck dissection when a recurrence is notproven.

Postchemoradiation Evaluation. We evaluatedthe accuracy of ultrasound-guided fine-needleaspiration cytology and found a sensitivity of80% but a specificity of only 42%. Ultrasound-guided fine-needle aspiration cytology used forthe N staging before treatment was reported toestablish an accuracy of 86% to 97% with highspecificity (83% to 100%).21–24,36 Knappe et al23

concluded that specificity was negatively influ-enced by misinterpretation of the smears by aless-experienced pathologist. When the N0-neckwas observed with ultrasound-guided fine-nee-dle aspiration cytology during follow-up, a sensi-tivity of 50% to 92% and a specificity that

approaches 100% was found.25,37–39 Consideringthe high specificity of ultrasound-guided fine-needle aspiration cytology when used for detec-tion of lymph node metastases in untreatednecks, it seems that the low specificity in thisstudy is attributable to the effects of chemora-diation. To our knowledge, no other reports onthis phenomenon are published.

The incidence of viable residual tumor cellsin 43% of the neck dissection specimens is inagreement with the reported incidences in theliterature, ranging from 29% to 56%.3,18,40–42

One should realize, however, that pathologistsonly look at a limited number of slides, with theconsequence of missing viable micrometasta-ses.43 Moreover, the clinical meaning of findingonly necrotic tumor cells is not known.

Increasing evidence exists that FDG-PETmight be a valuable modality to evaluate res-ponse of the neck after chemoradiation. Brkovichet al17 and Pordeddu et al44 found a high nega-tive predictive value of FDG-PET for diagnosticevaluation of the lymph nodes after chemoradia-tion. Results of PET are promising, but the accu-racy seems to be dependent on the intervalbetween treatment and imaging. A minimuminterval of 8 weeks seems advisable.44–47 FDG-PET-CT might increase this accuracy further.48

Selective or Modified Radical Neck Dissec-

tion. The type of neck dissection is another dif-ficult area of discussion. Although patientselection certainly played a role, in this seriesthe overall survival was significantly better forpatients with an SND (p ¼ .04) in univariateanalyses, but this lost significance in multivari-ate analysis (p ¼ .06). This is most likely theresult of a bias caused by pretreatment N classi-fication, although this was not statistically dif-ferent for both groups. In the majority of cases,an SND was performed for removal of 1 residualmass. Robbins et al20 also reported a better re-gional control and overall survival for patientswho underwent an SND compared with patientswho underwent an MRND. Recently hedescribed promising results of superselectiveneck dissection for patients with persistentnodal disease confined to 1 level.49 It seemstherefore not necessary to perform a comprehen-sive neck dissection in cases with limited resid-ual or recurrent disease. However, the observedtrend of better regional control after MRND (5-year regional control rate MRND 90%, SND77%, p ¼ .70) indicates that decisions for SND


should still be made with caution. When it isunclear which neck levels are involved, aMRND is recommended.

CONCLUSION

Considering the good regional control rate in thisstudy and the high rate of unnecessary neck dis-sections with a theoretically planned neck dissec-tion strategy, we conclude that a careful watch-and-wait strategy is worthwhile and safe.Regarding overall survival, patients with recur-rent disease seem to benefit more from this strat-egy than patients with residual disease.

Because the rate of residual or recurrent dis-ease is high in N2-3 necks, in these patients wenow perform a neck dissection in case of doubton the response in the neck at imaging or clini-cal examination. In a selected patient group itseems safe to perform a SND, although this de-cision should be made with caution. Comparedwith untreated necks, specificity of ultrasound-guided fine-needle aspiration cytology is signifi-cantly lower after chemoradiation, making thistechnique less reliable in these patients. Morereliable detection techniques of regional recur-rences are therefore needed.

REFERENCES

1. Pignon JP, Bourhis J, Domenge C, Designe L. Chemo-therapy added to locoregional treatment for head andneck squamous-cell carcinoma: three meta-analyses ofupdated individual data. MACH-NC CollaborativeGroup. Meta-analysis of chemotherapy on head and neckcancer. Lancet 2000;355:949–955.

2. Forastiere AA. Larynx preservation trials: a critical ap-praisal. Semin Radiat Oncol 1998;8:254–261.

3. McHam SA, Adelstein DJ, Rybicki LA, et al. Who meritsa neck dissection after definitive chemoradiotherapy forN2–N3 squamous cell head and neck cancer? Head Neck2003;25:791–798.

4. Narayan K, Crane CH, Kleid S, Hughes PJ, Peters LJ.Planned neck dissection as an adjunct to the manage-ment of patients with advanced neck disease treatedwith definitive radiotherapy: for some or for all? HeadNeck 1999;21:606–613.

5. Weisman RA, Robbins KT. Management of the neck inpatients with head and neck cancer treated by concurrentchemotherapy and radiation. Otolaryngol Clin North Am1998;31:773–784.

6. Robbins KT, Wong FS, Kumar P, et al. Efficacy of tar-geted chemoradiation and planned selective neck dissec-tion to control bulky nodal disease in advanced head andneck cancer. Arch Otolaryngol Head Neck Surg1999;125:670–675.

7. Corry J, Smith JG, Peters LJ. The concept of a plannedneck dissection is obsolete. Cancer J 2001;7:472–474.

8. Pellitteri PK, Ferlito A, Rinaldo A, et al. Planned neckdissection following chemoradiotherapy for advanced

head and neck cancer: is it necessary for all? Head Neck2006;28:166–175.

9. Grabenbauer GG, Rodel C, Ernst-Stecken A, et al. Neckdissection following radiochemotherapy of advancedhead and neck cancer–for selected cases only? RadiotherOncol 2003;66:57–63.

10. Lavertu P, Bonafede JP, Adelstein DJ, et al. Comparisonof surgical complications after organ-preservation ther-apy in patients with stage III or IV squamous cell headand neck cancer. Arch Otolaryngol Head Neck Surg1998;124:401–406.

11. Davidson BJ, Newkirk KA, Harter KW, Picken CA,Cullen KJ, Sessions RB. Complications from planned,posttreatment neck dissections. Arch Otolaryngol HeadNeck Surg 1999;125:401–405.

12. Sassler AM, Esclamado RM, Wolf GT. Surgery afterorgan preservation therapy. Analysis of wound complica-tions. Arch Otolaryngol Head Neck Surg 1995;121:162–165.

13. Sanguineti G, Corvo R, Benasso M, et al. Managementof the neck after alternating chemoradiotherapy foradvanced head and neck cancer. Head Neck 1999;21:223–228.

14. Wolf GT, Fisher SG. Effectiveness of salvage neck dissec-tion for advanced regional metastases when inductionchemotherapy and radiation are used for organ preserva-tion. Laryngoscope 1992;102:934–939.

15. Ojiri H, Mendenhall WM, Stringer SP, Johnson PL,Mancuso AA. Post-RT CT results as a predictive modelfor the necessity of planned post-RT neck dissection inpatients with cervical metastatic disease from squamouscell carcinoma. Int J Radiat Oncol Biol Phys 2002;52:420–428.

16. Ojiri H, Mancuso AA, Mendenhall WM, Stringer SP.Lymph nodes of patients with regional metastases fromhead and neck squamous cell carcinoma as a predictor ofpathologic outcome: size changes at CT before and afterradiation therapy. AJNR Am J Neuroradiol 2002;23:1627–1631.

17. Brkovich VS, Miller FR, Karnad AB, Hussey DH,McGuff HS, Otto RA. The role of positron emission to-mography scans in the management of the N-positiveneck in head and neck squamous cell carcinoma afterchemoradiotherapy. Laryngoscope 2006;116:855–858.

18. Stenson KM, Haraf DJ, Pelzer H, et al. The role of cervi-cal lymphadenectomy after aggressive concomitantchemoradiotherapy: the feasibility of selective neck dis-section. Arch Otolaryngol Head Neck Surg 2000;126:950–956.

19. Stenson KM, Huo D, Blair E, et al. Planned post-chemo-radiation neck dissection: significance of radiation dose.Laryngoscope 2006;116:33–36.

20. Robbins KT, Doweck I, Samant S, Vieira F. Effectivenessof superselective and selective neck dissection for ad-vanced nodal metastases after chemoradiation. ArchOtolaryngol Head Neck Surg 2005;131:965–969.

21. de Jong RJB, Rongen RJ, Verwoerd CD, et al. Ultra-sound-guided fine-needle aspiration biopsy of necknodes. Arch Otolaryngol Head Neck Surg 1991;117:402–404.

22. Hodder SC, Evans RM, Patton DW, Silvester KC. Ultra-sound and fine needle aspiration cytology in the stagingof neck lymph nodes in oral squamous cell carcinoma. BrJ Oral Maxillofac Surg 2000;38:430–436.

23. Knappe M, Louw M, Gregor RT. Ultrasonography-guidedfine-needle aspiration for the assessment of cervical me-tastases. Arch Otolaryngol Head Neck Surg 2000;126:1091–1096.

24. Takes RP, Knegt P, Manni JJ, et al. Regional metastasis inhead and neck squamous cell carcinoma: revised value ofUS with US-guided FNAB. Radiology 1996;198:819–823.


25. van den Brekel MW, Reitsma LC, Quak JJ, et al. Sono-graphically guided aspiration cytology of neck nodes forselection of treatment and follow-up in patients with N0head and neck cancer. AJNR Am J Neuroradiol 1999;20:1727–1731.

26. Balm AJ, Rasch CR, Schornagel JH, et al. High-dosesuperselective intra-arterial cisplatin and concomitantradiation (RADPLAT) for advanced head and neck can-cer. Head Neck 2004;26:485–493.

27. Robbins KT, Kumar P, Wong FS, et al. Targeted chemo-radiation for advanced head and neck cancer: analysis of213 patients. Head Neck 2000;22:687–693.

28. Robbins KT, Kumar P, Regine WF, et al. Efficacy of tar-geted supradose cisplatin and concomitant radiationtherapy for advanced head and neck cancer: the Mem-phis experience. Int J Radiat Oncol Biol Phys 1997;38:263–271.

29. Hoebers FJ, Heemsbergen W, Balm AJ, van Zanten M,Schornagel JH, Rasch CR. Concurrent chemoradia-tion with daily low dose cisplatin for advanced stagehead and neck carcinoma. Radiother Oncol 2007;85:42–47.

30. Lefebvre J, Horiot J, Rolland F, et al. Phase III study onlarynx preservation comparing induction chemotherapyand radiotherapy versus alternating chemoradiotherapyin resectable hypopharynx and larynx cancers. EORTCprotocol 24954–22950. J Clin Oncol 2007;25 (Suppl18):A-LBA6016–303s (abstract).

31. Gupta T, Agarwal JP. Planned neck dissection followingchemo-radiotherapy in advanced HNSCC. Int SeminSurg Oncol 2004;1:6.

32. Goguen LA, Posner MR, Tishler RB, et al. Examiningthe need for neck dissection in the era of chemoradiationtherapy for advanced head and neck cancer. Arch Otolar-yngol Head Neck Surg 2006;132:526–531.

33. Mendenhall WM, Villaret DB, Amdur RJ, Hinerman RW,Mancuso AA. Planned neck dissection after definitiveradiotherapy for squamous cell carcinoma of the headand neck. Head Neck 2002;24:1012–1018.

34. Argiris A, Stenson KM, Brockstein BE, et al. Neck dis-section in the combined-modality therapy of patientswith locoregionally advanced head and neck cancer.Head Neck 2004;26:447–455.

35. Richey LM, Shores CG, George J, et al. The effectivenessof salvage surgery after the failure of primary concomi-tant chemoradiation in head and neck cancer. Otolaryn-gol Head Neck Surg 2007;136:98–103.

36. van den Brekel MWM, Castelijns JA. Radiologic evalua-tion of neck metastases: the otolaryngologist’s perspec-tive. Semin US-CT-MRI 1999;20:162–174.

37. Righi PD, Kopecky KK, Caldemeyer KS, Ball VA, Weis-berger EC, Radpour S. Comparison of ultrasound-fineneedle aspiration and computed tomography in patientsundergoing elective neck dissection. Head Neck 1997;19:604–610.

38. van den Brekel MW, Castelijns JA, Reitsma LC, Lee-mans CR, van der Waal I, Snow GB. Outcome of observ-ing the N0 neck using ultrasonographic-guided cytologyfor follow-up. Arch Otolaryngol Head Neck Surg 1999;125:153–156.

39. Yuasa K, Kawazu T, Kunitake N, et al. Sonography forthe detection of cervical lymph node metastases amongpatients with tongue cancer: criteria for early detectionand assessment of follow-up examination intervals.AJNR Am J Neuroradiol 2000;21:1127–1132.

40. Budach VGDS, Haake K, Stuschke M, et al. Acceleratedchemoradiation to 70.6 Gy is more effective than acceler-ated radiation to 77.6 Gy alone—two year results of aGerman multicenter randomized trial. Int J RadiatOncol Biol Phys 2000;48 (Suppl 1):150–151 (abstract).

41. Calais G, Alfonsi M, Bardet E, et al. Stage III and IVcancers of the oropharynx: results of a randomized studyof Gortec comparing radiotherapy alone with concomi-tant chemotherapy [in French]. Bull Cancer 2000;87Spec No:48–53.

42. Velazquez RA, McGuff HS, Sycamore D, Miller FR. Therole of computed tomographic scans in the managementof the N-positive neck in head and neck squamous cellcarcinoma after chemoradiotherapy. Arch OtolaryngolHead Neck Surg 2004;130:74–77.

43. van den Brekel MW, Stel HV, van derValk P, van derWaal I, Meyer CJ, Snow GB. Micrometastases fromsquamous cell carcinoma in neck dissection specimens.Eur Arch Otorhinolaryngol 1992;249:349–353.

44. Porceddu SV, Jarmolowski E, Hicks RJ, et al. Utility ofpositron emission tomography for the detection of dis-ease in residual neck nodes after (chemo)radiotherapy inhead and neck cancer. Head Neck 2005;27:175–181.

45. Rogers JW, Greven KM, McGuirt WF, et al. Can post-RTneck dissection be omitted for patients with head-and-neck cancer who have a negative PET scan after defini-tive radiation therapy? Int J Radiat Oncol Biol Phys2004;58:694–697.

46. McCollum AD, Burrell SC, Haddad RI, et al. Positronemission tomography with 18F-fluorodeoxyglucose topredict pathologic response after induction chemother-apy and definitive chemoradiotherapy in head and neckcancer. Head Neck 2004;26:890–896.

47. Yao M, Smith RB, Graham MM, et al. The role of FDGPET in management of neck metastasis from head-and-neck cancer after definitive radiation treatment. Int JRadiat Oncol Biol Phys 2005;63:991–999.

48. Nayak JV, Walvekar RR, Andrade RS, et al. Deferringplanned neck dissection following chemoradiation forstage IV head and neck cancer: the utility of PET-CT.Laryngoscope 2007;117:2129–2134.

49. Robbins KT, Shannon K, Vieira F. Superselective neckdissection after chemoradiation: feasibility based on clin-ical and pathologic comparisons. Arch Otolaryngol HeadNeck Surg 2007;133:486–489.


Effectiveness of salvage selective and modified radical neck dissection for regional pathologic...

Documents

Transcript of Effectiveness of salvage selective and modified radical neck dissection for regional pathologic...