Is Modified Radical Neck Dissection Only A Staging Procedure?

CHRISTOPHER J. O'BRIEN, FRACS," SENGJAW SOONG, PHD,t MARSHALL M. URIST, MD.+ AND WILLIAM A. MADDOX, MD$

This retrospective study evaluates the clinical benefit of modified radical neck dissection among patients with squamous carcinoma of the upper aerodigestive tract. Ninety-eight modified neck dissections were performed in 86 patients over a 5-year period. The procedure entailed removal of the submaxillary and jugular chain nodes while the posterior triangle was not dissected. Thirty-two patients received postoperative radiotherapy. Lymph nodes were histologically positive in 55 of 98 dissections (56%).Among 72 determinate patients, recurrence in the dissected neck occurred in 8 of 38 with positive nodes and none of 34 with negative nodes (P c 0.05). These recurrences occurred in patients who had clinically palpable nodes preoperatively. Postoperative radiotherapy did not significantly alter the overall recurrence rate or survival of patients with positive nodes. Cumulative disease-free survival at 5 years was 70%overall. It is concluded that the modified neck dissection described is appropriate in the clinically negative neck or when regional disease is early ( k , N1) and located in the submandibular triangle. Postoperative radiotherapy should be given if more than one node is involved histologically or if extracapsular spread is present. Cancer 59:994-999, 1987.

now widely performed and although techniques vary in name and detail, they share fundamental similarities. In each procedure the aim is to remove the lymphatics of the neck while attempting to limit morbidity by sparing certain anatomical structures. These are, principally, the sternocleidomastoidmuscle (SCM), the spinal accessory nerve (SAN), and the internaljugular vein (IJV). Modified radical neck dissection is especially useful when carried out as an elective procedure in the clinically negative neck because it is associated with less cosmetic and functional disability than classical radical neck dissection. However, the oncologic effectiveness of MRND in patients with epidermoid carcinoma of the upper aerodigestive tract is not yet clear. Authors are divided in their opinions as to when a modified procedure should be car-

ODIFIED RADICAL NECK DISSECTIONS (MRND) are

ried out in the clinically positive neck and when postoperative radiation therapy should be used. In an attempt to more clearly define the role and efficacy of MRND, the experience of two of the authors (W.A.M and M.M.U) has been reviewed. Patients and Methods Modified radical neck dissection was first carried out at the University of Alabama Hospitals in 1979, and by December 1984, a total of 98 operations had been performed in 86 patients. These procedures were unilateral in 74 patients and synchronous bilateral operations in 12. In general, patients undergoing MRND had either (1) clinically normal necks and at least a 30% risk of occult nodal metastases based on the T-stage and histologic characteristics, or (2) clinically positive cervical nodes that were not in proximity to the course of the spinal accessory nerve. The operative technique, which was developed by one of the authors (W.A.M), has been described elsewhere.' The procedure consists of dissection of the submandibular triangle and upper, middle, and lower jugular chains, sparing the SCM, SAN, and cervical plexus (CP) but removing the IJV. The posterior triangle is not dissected. When synchronous bilateral procedures were performed, the IJV on the side opposite the primary (or the clinically less involved side) was spared. Postoperative radiation therapy was given to 32 patients. This was commenced within 4 to 6 weeks of surgery. Treatment doses

From the Section of Surgical Oncology, Department of Surgery and the Section of Biostatistics, Comprehensive Cancer Center, University of Alabama in Birmingham, Birmingham, Alabama. Supported by grants from the NCI CA382 15. * Fellow in Head and Neck Oncology, Section of Surgical Oncology. t Chief, Section of Biostatistin, ComprehensiveCancer Center. $ Chief, Section of Surgical Oncology, Department of Surgery. Q Clinical Professor, Section of Surgical Oncology, Department of Surgery. Address for reprints: Marshall M. Urist, MD, University Station, 320 Kracke Building, Birmingham, AL 35294. The authors thank Ms.Judy Smith and Ms. Judy Warren for assisting with this study and Ms. Michelle Dunn for typing the manuscript. Accepted for publication October 1, 1986.

994

No. 5

MODIFIED RADICAL NECKDISSECTION

O'Brien et al.

995

ranged from 5000 to 6000 rad given over 5 to 6 weeks. Initially, radiation fields encompassed the primary site and the ipsilateral neck but more recently the contralateral neck has been included in the treatment schedule. Seventy-two of the 86 patients were determinate because they had been followed up for a minimum of 2 years or developed recurrence of their disease. Three patients who were followed for less than 2 years and l l who died of other causes within 2 years of MRND were regarded as being indeterminate. Differences in recurrence rates in the ipsilateral neck, with disease controlled at the primary site, were analyzed using a x2 test with Yates' correction factor. Disease-free survival was calculated by the method of Kaplan and Meier and patients alive with disease were counted as deaths. Differences between survival rates were compared by a generalized Wilcoxon test.
Results

No. Patients

35 301
25

,,
27
40

2,

"30

50

60

70

80

90

Age( years) Females Males

FIG. 1. Age/sex distributionof 86 patients undergoing MRND.

Among the 86 patients, there were 65 men and 21 women with ages ranging from 33 to 87 years (median, 58 years). The age/sex distribution is shown in Figure 1. Table 1 shows the sites of primary tumors involved along with clinical node status for tumors at each site. This table, therefore, reflects the indications for MRND. For example, 10 of the 1 1 MRNDs for lip carcinoma were therapeutic procedures. Neck dissections were elective in 52 of 86 patients (60%)and therapeutic in 34 (40%).Table 2 shows the clinical stage of patients before MRND, according to the guidelines of the American Joint ComO primaries were both mittee on Cancer ( 1983). The T 1N hypopharyngeal cancers. Patients who had treatment for recurrent disease at the primary site along with MRND, were staged Tx for the primary. Those with previously treated primaries and no evidence of recurrence at the primary site but with clinical disease in the neck were staged TO. One patient with an unknown primary site was also staged TO. Histologic Node Involvement At the time of pathologic analysis, the total yield of lymph nodes from neck dissection specimens and the numbers of involved nodes were noted. This information is represented in Figures 2 and 3. Lymph nodes were histologically positive in 55 of 98 neck dissections (56%). Table 3 shows the correlation of the clinical and histologic findings. The false negative rate of clinical evaluation of the neck was 37% (22 of 60) and the false positive rate was 13%(5 of 38). Recurrences Table 4 shows the recurrence data of patients subgrouped according to the number of histologically

positive nodes found. The designated site of recurrence was the most proximal site at which disease recurred. Therefore the ipsilateral neck was called the recurrence site only when disease was controlled at the primary site. Among the 12 patients who had synchronous bilateral MRNDs, 7 had bilateral disease proven histologically and they form a separate group; 3 had no histologic disease and are included among the 38 patients with no positive nodes; and 2 with unilateral disease histologically were included with other patients according to the number of lymph nodes involved.
TABLE 1 . Site of Primary Tumor and Clinical Node Status of 86 Patients Having MRND Site Lip Oral cavity Tongue Floor of mouth Buccal mucosa Retromolar trigone Alveolar ridge Oropharynx Tonsil Base of tongue Soft palate Larynx Supraglottis Glottis Hypopharynx Unknown primary Total Nodes + VE
10

Nodes - VE
1

Total
11

2 4 5 0 6
1

6
11
1

8
15

4 1 0

6 4 16

0 1
1

2 2 0
6 2 I
0

3
2
1

2
1 1

I 4 8
1

34

52

86

+ V E positive; - V E negative; MRND: modified radical neck dissection.

996

CANCER March 1 1987

VOl. 59

TABLE 2. TNM Stage for 86 Primary Tumors at the Time of MRND NO N1 2

11

N2
0 3 0 2 1 0 6

N3

Total 3 15 5 43 18 2 86

TX T O T1 T2 T3 T4
Total

1 0 2 36 13 0

0
1
1 1

2 4 4 0 23

0 2
5

52

MRND: modified radical neck dissection.

Where lymph nodes were histologically negative 4 of 34 determinate patients (12%) developed recurrence but in no instance did this involve the ipsilateral neck. Where nodes were histologically positive, recurrence developed in 15 of 38 patients (39%)and this involved the ipsilateral neck in 8 patients (21%). The presence of histologically positive nodes was associated with a significantlyincreased recurrence rate at all sites (P < 0.05) and, in particular, in the ipsilateral neck (P < 0.05). Although the numbers are small, the overall incidence of recurrence increased with the number of lymph nodes involved. Only 1 of 12 determinate patients with one positive node developed ipsilateral neck recurrence and this patient had extracapsular invasion by nodal metastatic disease. The others were mainly patients with lip cancers and a single node in the submandibular triangle. Postoperative Radiotherapy Twenty-eight patients with histologically positive nodes had postoperative radiation therapy to the ipsilateral neck. Twenty-two of these patients are determinate and 8 of these developed recurrence (36%). Recurrence in the ipsilateral neck occurred in four patients in this group ( 18%) whereas three failed at the primary site and one had distant metastases. Two others developed new primary cancers. The four patients who recurred in the neck all had clinically palpable nodes preoperatively. Three of these had histologic evidence of extracapsular spread and more than

four nodes involved while in the other patient, the recurrence occurred in the posterior triangle and review of radiation fields showed that this area had not been irradiated postoperatively. Twenty-one patients with positive nodes were not irradiated. Sixteen of these patients are determinate and 7 recurred (44%). The recurrence rate in the ipsilateral neck in this group was 25% (4 of 16) and all these patients had clinically palpable nodes preoperatively. Transcapsular spread of tumor was present in each patient. Of the remaining three patients, two failed at the primary site and one at a distant site. Therefore the recurrence rates overall, and in the ipsilateral neck, were not significantlydifferent for irradiated and nonirradiated patients with positive nodes although irradiated patients tended to have worse disease. Furthermore, all recurrences in the ipsilateral neck occurred in patients who had clinically palpable nodes preoperatively and extracapsular spread was present in each case. Survival Using the direct method of survival analysis (the number of patients alive and well compared with the total number of determinate patients) the disease-free survival rate at 2 years was 78% (56 of 72 patients). Thirteen patients ( 18%)had died of their disease and 3 patients were alive with disease. The estimated cumulative disease-free survival rate at 5 years by the method of Kaplan and Meier was 70%.The survival of patients without histologic evidence of lymph node metastases was significantlybetter than that of patients with nodal metastases (Fig. 4). Irradiated and nonirradiated patients with positive nodes had similar survival rates (Fig. 5). Disease-free survival decreased as the number of positive nodes increased. Patients with no positive nodes had significantly better survival than patients with two to four and more than four nodes (Fig. 6). Among patients with histologically positive nodes there was a survival advantage at 2 years for patients who were clinically negative (83%) compared with those whose nodes were clinically positive (56%). However, by 5 years the survival rates approached each other (Fig. 7). Complications

351

No. o f Neck Dissections

304

31

15 10

.no jo j0 j0 5ho
3

7 0 0

No. of Nodes
FIG.2. Frequency histogram showing the yield of lymph nodes in 94 MRNDs. In 3 I cases between 20 and 29 nodes were found.

Two patients died in the perioperative period. One was an 80-year-old woman who died of a myocardial infarction. The other was a 63-year-old man who died of carotid artery rupture related to a pharyngocutaneous fistula after laryngectomy. Eight other patients suffered complications directly attributable to the neck dissection. There were three chylous fistulae (one of which required surgical closure), one wound infection, and one case of necrosis of the skin edges of the neck wound. Three other patients had nerve injuries: two had vagal neuropraxias which caused temporary hoarseness, and one patient suffered

No. 5

MODIFIED RADICAL NECK DISSECTION OBrien et al.

997

TABLE 3. Correlation of Clinical and Histologic Examination for 98 MRNDs

No. of Neck Dissections

40

20 15
5 10

1
0

Clinically positive Histologically positive Histologically negative

33 5 38

Clinically negative Total 22 4 n 4 r blr y y M 55

38
60

43

26

n
23
U

Total

98

MRNDs: modified radical neck dissections.

2-4

>4

NO. Nodes Involved

FIG. 3. Frequency histogram showing number of histologicallypositive nodes in each neck dissection.

hemidiaphragmatic paresis from inadvertent division of the left phrenic nerve.

Discussion

The management of the neck in patients with squamous carcinomas of the upper aerodigestive tract varies according to the clinical circumstances. Clinically palpable nodes are usually treated by classical radical neck dissection, and radiation therapy often is added in an attempt to prevent subsequent recurrence in the neck. Recently however, patients with clinically positive neck nodes have been treated by modified radical dissection, but the place of this operation has not been fully defined. Management of the clinically negative neck is even more controversial. Opinion is divided as to whether the best treatment is elective irradiation, elective neck dissection, or observation with subsequent therapeutic neck dissection when clinical metastases appear. In fact, any of these options may be appropriate in an individual patient, depending on the site and stage of the primary tumor. At this time,

if an elective neck dissection is planned, there appears to be little justification for carrying out a radical operation. Judicious modified radical neck dissection could identify patients with occult metastases, with low morbidity, and obviate unnecessary radical neck dissection or elective irradiation in many cases. In this retrospective study, modified neck dissections have been carried out as both therapeutic and elective procedures. Elective removal of occult nodal metastases was associated with a significantly better survival at 2 years than when nodes were clinically positive. However, the difference in survival of the two groups had all but disappeared by 5 years. The implication is that removal of involved lymph nodes before they became clinically evident has little effect on ultimate survival and this is the argument advanced by antagonists of elective neck dissection. However, in this nonrandomized series no definite conclusions can be made about the potential survival benefit from elective MRND. Bocca et al. recently reported their series of 1500 functional neck dissections of which 1200 were elective procedures.* However, 87% of their patients had carcinoma of the larynx and neither the distribution of cancers within the larynx nor the incidence of histologically positive nodes in these clinically negative necks were stated. Since glottic carcinoma accounts for 50% to 75% of laryngeal cancers and rarely metastasizes to the neck unless a d ~ a n c e d ,it ~is -~ not clear just how many patients needed or benefited from elective neck dissection. In the same series, 17 1 patients with clinical N 1 or N2 disease had a

TABLE 4. Initial Site of Recurrence According to Number of Nodes Involved Site of recurrences Nodes positive Total patients Determinate patients Primary Ipsilat neck Contralat neck Distant Total
4 (12%) 1(8%) 5 (31%) 4 ( 100%) 5 (83%)

0
1

2-4 >4 Bilateral nodes

38 18 18
5 7

34 12
16 4

6 12

3 0 3 0 2 8

0
1 1

4 2

I 0 0 0

0 0
1

0
1

0
1

Total

86

19 (26%)

Ipsilat: ipsilateral; contralat: contralateral.

998

CANCER March 1 1987

Vol. 59

6
0
C

90

80 70 60
50
40

.g
v)

c .

0 .-

30
P<O.Ol

* 20
O s
10

0
6
12 18 24

30

36

42

48

54

60

Time (months)
FIG. 4. Cumulative disease-free survival curves showing that patients without positive nodes had significantly better survival than patients with positive nodes.

therapeutic modified radical neck dissection without postoperative irradiation and 30%recurred in the ipsilatera1 neck. This suggests that the procedure is not ideal when nodes are clinically palpable, although the authors believed that the results compared favorably with a group of historical controls treated by classical radical neck dissection. In the current series, all eight recurrences in the ipsilateral neck occurred after MRND in patients with clinically positive nodes, again highhghting the limitations of therapeutic MRND. Other retrospective studies comparing patients treated by modified neck dissection with those treated by radical neck dissection also have con-

With

R T

(n 28)

.-

Wilhoul

R T (n=20)

3
c
v)

50
40

.Q) c

30

a
. o

20
10
I
I I

. c

0 &

p 0.48
I
I

12

18

24

30

36

42

48

54

60

Time (months)
R.T. = Radiation Therapy

FIG. 5. The survival of patients with positive nodes was not significantly improved by the addition of postoperative radiotherapy.

cluded that MRND is appropriate only when clinical disease in the neck is early. Lingeman et al. suggested that conservative neck dissection be used for subclinical or N 1 disease whereas Molinari et aL8believe this procedure is safe when palpable nodes are nonsuspect or less than 2.5 cm in diameter and mobile. In fairness however, any retrospective comparisons are invalid because of uncontrolled allocation of patients to treatment groups. Histologic factors which increase the likelihood of neck recurrence after MRND apply equally to radical neck dissection. These are the presence of extracapsular invasion and the involvement of multiple nodes at multiple level^.^-'^ In a series of 967 patients recently reported by Byers from the M. D. Anderson Hospital, the addition of radiation therapy to MRND among patients with these characteristics was shown to reduce recurrence in the ipsilateral neck from 26% to 13%.13A recent study from our own institution also indicated that the addition of radiotherapy to neck dissection halved the incidence of ipsilateral neck recurrence in node positive patients, although overall survival was not improved.l4 In the current series, the finding of histologicallypositive nodes in MRND specimens was associated with a significantly increased risk of recurrence in the ipsilateral neck. The recurrence rate at all sites also increased with the number of lymph nodes histologically involved. This latter association is not necessarily causal but may reflect the fact that more aggressive tumors tend to involve more lymph nodes as well as spread to other sites. Among determinate patients with histologically negative nodes, no recurrences occurred in the ipsilateral neck and there were no deaths from disease, although one patient is alive with disease. Also, when only one lymph node was histologically positive, failure in the ipsilateral neck occurred only when extracapsular spread was present (one patient). Modified radical neck dissection appeared efficacious for early histologic disease in the submandibular triangle. Recurrence in the ipsilateral neck occurred among patients with clinically palpable nodes when two or more lymph nodes were involved histologically, but extracapsular spread was present in each case. The addition of postoperative radiation therapy did not appear to significantly alter the rate or pattern of recurrence among patients with histologically positive nodes, but the irradiated patients in general had a larger number of involved lymph nodes than those who recurred without radiation therapy, indicating that they were at greater risk. One patient developed recurrence in the posterior triangle. This patient was clinically staged N2 preoperatively, because two nodes were palpable in the submandibular triangle, and had neither dissection nor irradiation of the posterior triangle. Mantravadi et al. also found an increased risk of recurrence in the posterior triangle when postoperative radiotherapy did not include this area.l 5

No. 5

MODIFIED RADICAL NECK DISSECTION O'Brien et al.

a
. c

999

Thus when postoperative radiation therapy is to be added to the modified neck dissection described in this series, it should encompass the entire ipsilateral neck to decrease the risk of recurrence in the posterior triangle. It appears, therefore, that modified radical neck dissection is primarily a staging procedure. When carried out electively, the morbidity is low and patients with occult metastatic disease can be identified. In this series, 38% of necks clinically judged to be negative contained occult disease. If the yield of positive lymph nodes is none or one, without extracapsular spread, then no further treatment to the neck is necessary since recurrence is unlikely to occur. If, however, more than one node is positive or extracapsular spread is present, then postoperative radiation therapy should be added to the modified neck dissection. Although our results failed to show any definite benefit from postoperative radiation therapy, its advantages in improving regional control have been r e p ~ r t e d ' ~and - ' ~ we continue to adhere to this principle. Modified neck dissection alone can effectively control disease in the neck only in patients with minimal histologic metastatic disease. Otherwise, its efficacy is dependent upon its combination with planned postoperative radiotherapy. Nonetheless this procedure has distinct advantages over classical radical neck dissection as an elective operation and also may prevent a number of patients undergoing unnecessary elective neck irradiation. Whether or not MRND can be used effectively instead of radical neck dissection among more patients with palpable nodes can only be determined by a randomized clinical trial in which both procedures are combined with adjuvant radiotherapy according to the pathologic findings.

In .-

100 go

80

1 Clinically

ve ln-201

a 70
.? 60
50
v)

'

Clinically

+ye

(n 281

0'

I I I

12

18

24

30

36

42

48

54

60

Time (months)
FIG. 7. Among patients with histologically positive nodes, those with clinically negative disease had better survival than patients with clinically positive disease at 2 years, but by 5 years there was little difference.

REFERENCES
1. OBrien CJ, Urist MM, Maddox WA. Modified radical neck dissection: Terminology, technique and indications. A m J Surg (in press). 2 . Bocca E, Pignataro 0,Oldini C, Cappa C. Functional neck dissection: An evaluation and review of 843 cases. Laryngoscope 1984;94:

942-945. 3 . Smith RR, Caulk RM, Russell WO, Jackson CL. End results in
600 laryngeal cancers using the American Joint Committee's proposed method of stage classification and end results reporting. Surg Gynecol Obstet 196 1 ; I 13:435-444. 4 . Vermund H.Role of radiotherapy in cancer of the larynx as related to the TNM system of staging. Cancer 1970;25:485-504. 5 . Harwood AR. Cancer of the larynx: The Toronto experience. J Otolaryngol 1982;(Suppl)l1:1-21. 6. Ogura JH, Biller HF, Wette R. Elective neck dissection for pharyngeal and laryngeal cancers: An evaluation. Ann Otol Rhino1 Laryngol

1971;80:646-651. 7. Lingeman RE, Helmus C, Stephens R, Ulm J. Neck dissection: Radical or conservative. Ann Otol 1977;86:737-744. 8. Molinari R, Cantu G , Chiesa F, Grandi C. Retrospective comparison of conservative and radical neck dissection in laryngeal cancer. Ann Otol 1980;89578-581. 9. Shah JP, Cendon RA, Farr HW, Strong EW. Carcinoma of the oral cavity: Factors affecting treatment failure at the primary site and neck. Am JSurg 1976; 132504-507. 1 0 . Kalnins IK, Leonard AG, Sako K, Razack MS, Shedd DP. Correlation between prognosis and degree of lymph node involvement in carcinoma of the oral cavity. Am J Surg 1977; 134450-454. 1 I . Johnson JT, Barnes EL, Myers EN, Schramm VL, Broochovitz D, Sigler BA. The extracapsular spread of tumors in cervical node metastasis. Arch Otolaryngol 1981;107:725-729. 12. Carter RL, Barr L, OBrien CJ, So0 KC, Shaw HJ. Transcapsular spread of metastatic squamous carcinoma from cervical lymph nodes. Am JSurg 1985;150495-499. 13. Byers RM.Modified neck dissection: A study of 967 cases from 1970 to 1980.Am JSurg 1985;150:414-421. 14. O'Brien CJ, Smith JW, Soong S-J, Unst MM, Maddox WA. Neck dissection with and without radiotherapy: Prognostic factors, patterns of recurrence and survival. Am JSurg 1986; 152:456-463. 15. Mantravadi RV, Skolnik EM, Haas RE, Applebaum EL. Patterns of cancer recurrence in the postoperatively irradiated neck. Arch Otolaryngol 1983;109:753-756. 16. Pearlman NW, Johnson FB, Kennaugh RC. Modified radical neck dissection and postoperative radiotherapy in squamous cell head and neck cancer. Am JSurg 1985; 150:488-490.

1. 1 Node I n = l 8 )

1 vs. 2 2 vs. 3 1 vs. 3 1.2.3 vs. 4

p=o.18 p=0.62 ~(0.05 p<O.Ol

20 10

01

12

18

24

30

36

42

48

54

60

Time (months)
FIG. 6.Cumulative disease-free survival according to the number of histologically positive nodes. Patients with bilateral positive nodes have been grouped with patients with more than four nodes.