Results of Salvage Treatment of the Neck in Patients With Oral Cancer

Background  About 50% of the patients with neck recurrences after the treatment of oral squamous cell carcinoma are not considered candidates for further treatment, and reported survival is generally poor.

Objective  To evaluate the prognostic importance of neck recurrences and results of salvage treatment in patients with oral carcinoma.

Patients  Five hundred thirteen patients with squamous cell carcinoma of the oral cavity underwent surgical treatment, with follow-up from less than 2 to 119 months (mean, 16.9 months).

Setting  Referral center, private or institutional practice, and ambulatory and hospital care center.

Intervention  Four hundred forty-eight patients underwent neck dissection, and 65, resection of the primary tumor only. Postoperative radiotherapy was used for 228.

Main Outcome Measures  Rates of neck recurrences and survival after salvage treatment.

Results  Eighty-two patients (16.0%) had neck recurrences, including ipsilateral in 44, contralateral in 31, and bilateral in 7. Most neck recurrences (77 [94%]) were diagnosed within 2 years. Salvage treatment was attempted in 51 patients (62%). Of the patients with a previously untreated side of the neck, 27 underwent radical neck dissections (11 ipsilateral and 16 contralateral) and only 5 remained with no evidence of disease. The significant factors associated with survival after neck recurrence were type of previous neck dissection (P<.001), previous postoperative radiotherapy (P = .003), and interval free of neck recurrence (P<.001).

Conclusions  Patients undergoing previous neck dissection and with recurrences diagnosed after 6 months are not usually candidates for curative salvage treatment and are at a high risk for death. Only 5 of 46 patients with recurrences in a previously untreated side of the neck survived after salvage treatment. Patients with neck recurrences have a poor prognosis, despite salvage treatment.

STATE-OF-THE-ART treatment of patients with oral cancer and N1 to N3 regional lymph node metastases consists of neck dissection (radical or selective in selected cases) and postoperative radiotherapy. The high incidence rate of occult cervical metastases in tumors of the oral cavity is a strong argument in favor of elective treatment of the neck. Implications of elective treatment include the advantage of treating subclinical disease, but treatment may lead to increased morbidity and cost that ideally should be avoided in patients with pN0 neck disease. Despite advances in multimodality treatment, local and neck recurrences remain the most frequent in patients with oral cancer. Of the patients who present with such recurrences, about 50% are not considered for further treatment because of the advanced disease stage, and reported survival is generally poor.^1-8

The need to evaluate the prognostic importance of initial therapy, survival free of neck recurrence, and salvage treatment of patients with treatment failures in the neck prompted this study. We analyzed a large series of patients from a single institution who underwent surgery (with and without neck dissection) to give an appropriate perspective of the results that can be achieved with the treatment of patients with this particular site of recurrence.

We retrospectively analyzed 513 consecutive patients with previously untreated squamous cell carcinoma of the oral cavity. Primary sites included the oral tongue (269 [52.4%]), floor of the mouth (135 [26.3%]), inferior gingiva (44 [8.6%]), and retromolar triangle (65 [12.7%]). All patients initially underwent curative surgery or surgery with postoperative radiotherapy at the Head and Neck Surgery and Otorhinolaryngology Department of the Centro de Tratamento e Pesquisa Hospital do Cancer A.C. Camargo, São Paulo, Brazil, from March 22, 1970, through December 28, 1992. With regard to sex, 437 patients (85.2%) were men and 76 (14.8%) were women, with ages ranging from 24 to 95 years (mean age, 56.4 years). As for ethnic groups, 449 (87.5%) were white, and 64 (12.5%) were nonwhite.

A total of 250 patients (48.7%) presented with clinically metastatic ipsilateral neck lymph nodes. In 28 patients (5.5%), the clinically metastatic lymph nodes were bilateral. Based on the clinical staging system proposed by the TNM Committee of the International Union Against Cancer,⁹ cancer in the 513 patients was classified as follows: stage I in 63 (12.3%); stage II, 120 (23.4%); stage III, 173 (33.7%); and stage IV, 157 (30.6%).

A total of 448 patients (87.3%) underwent a neck dissection. Of these, 426 procedures were continuous and 22 were discontinuous; 333 were unilateral and 115 were bilateral. Indications for neck dissections were elective (absence of clinically metastatic lymph nodes) or therapeutic (presence of clinically metastatic lymph nodes), resulting in 161 elective ipsilateral, 37 elective bilateral, 172 ipsilateral therapeutic, 50 therapeutic ipsilateral and elective contralateral, and 28 bilateral therapeutic procedures. The types of ipsilateral neck dissection used in each case included supraomohyoid (n = 46), classic radical (n = 319), modified radical with preservation of the accessory nerve (n = 76), and modified radical with preservation of the accessory nerve and the internal jugular vein (n = 7). The following procedures were used for treatment of the neck's contralateral side: supraomohyoid (n = 79), modified radical with preservation of the accessory nerve (n = 14), and modified radical with preservation of the accessory nerve and the internal jugular vein in (n = 22).

Postoperative radiotherapy was used for 228 patients (44.4%). The interval from surgery to the onset of radiotherapy ranged from 1 to 34 weeks (median, 5.7 weeks). During the study, indications varied broadly because of the characteristics of the primary tumor (close or involved surgical margins and presence of vascular embolization or perineural infiltration) or the presence of regional lymph node metastases (independent of the number or stage of involved lymph nodes or presence of transcapsular spread). In 12 patients undergoing postoperative radiotherapy, none of these characteristics were found. On the other hand, in 183 of the 285 patients not undergoing radiotherapy, at least 1 of those factors was present (principally close margins, vascular embolization, and perineural infiltration). The radiation source was a 4-mV linear accelerator in 171 patients and cobalt beam in 57.

Neck metastases were diagnosed in the surgical specimens of 223 (49.8%) of the 448 patients initially undergoing unilateral or bilateral neck dissections. Ipsilateral metastases were found in 182 cases; contralateral, in 5; and bilateral, in 36. Of all 448 patients undergoing neck dissection, metastatic lymph nodes were diagnosed in 218 (48.7%). Of the 115 patients undergoing simultaneous contralateral neck dissection, 41 (35.7%) presented with metastatic lymph nodes. Most of the 238 metastatic lymph nodes were located only at levels 1, 2, or 3; 198 were ipsilateral and 40 contralateral. Lymph nodes at levels 4 and 5 were identified at the ipsilateral side of the neck in 20 patients and at the contralateral side in 1 patient.

We measured the interval free of neck recurrence as the time in months elapsed from initial surgery to detection of neck recurrence. For survival after recurrence (SAR), we calculated the time in months from the salvage treatment (or recurrence detection for untreated patients) until death or the date of last follow-up information. The estimates of SAR were considered uncensored for all the cases in which death was the outcome. The Kaplan-Meier method was used to calculate the actuarial survival.¹⁰ Comparisons between distributions of survival rates among the categories of a single variable were performed according to the Mantel-Cox test.

During follow-up ranging from less than 2 to 119 months (mean, 16.9 months), 214 patients (41.7%) presented with recurrences and/or metastases, of which 82 (16.0%) were neck recurrences that were diagnosed from 1 to 83 months after initial surgical treatment (77 recurrences [94%] within 24 months). The distribution of all recurrences by site is presented in the following tabulation:

Of the 51 ipsilateral neck recurrences (including the 7 patients with bilateral recurrences), 32 (7.1%) were found in the 448 patients previously undergoing neck dissection and 19 (29.2%) among the 65 patients with no initial treatment of the neck. Contralateral neck recurrences were diagnosed in 38 of the 115 patients undergoing simultaneous contralateral neck dissection (including the 7 patients with bilateral recurrences), of whom 5 (4.3%) had initially undergone bilateral neck dissection. There were 33 (8.3%) recurrences in the 398 patients not initially undergoing a contralateral neck dissection. Most neck metastases and recurrences involved ipsilateral or contralateral levels 1 to 3 lymph nodes (Table 1). Postoperative radiotherapy did not influence significantly the rates of ipsilateral or contralateral neck recurrences (Figure 1).

Thirty-one (37.8%) of the 82 patients with neck recurrences were treated with chemotherapy with palliative intent or were considered not to qualify for any kind of oncologic treatment (palliative care only). The remaining patients were treated with palliative or curative intent. Radiotherapy was used as a single treatment in 15 patients (18.3%), and radiation doses varied from 3000 to 7000 rad (30-70 Gy) (median, 5000 rad [50 Gy]). Thirty-six patients with resectable neck recurrences underwent surgery, including surgery alone in 14 (17.1%) and surgery with postoperative radiotherapy in 22 (26.8%). Surgical procedures included the following: wide resection of the ipsilateral neck recurrence in 5 patients (associated with radiotherapy in 3), supraomohyoid dissection in 2 patients (bilateral in 1 patient, associated with radiotherapy), classic or modified radical neck dissection in 11 patients (ipsilateral in 3 and contralateral in 8), and classic or modified radical neck dissection associated with postoperative radiotherapy in 18 patients (ipsilateral in 8, contralateral in 8, and bilateral in 2).

Duration of survival after diagnosis of neck recurrences ranged from less than 1 to 145 months (mean, 20.2 months; median, 7.6 months). Survival of the 44 patients with ipsilateral metastases ranged from less than 1 to 145 months (mean, 18.4 months), whereas for the 31 patients with contralateral metastases, it ranged from less than 2 to 123 months (mean, 25.6 months), and in the 7 patients with bilateral metastases, from less than 2 to 24 months (mean, 8.0 months) (Figure 2). Of the 31 patients who did not undergo salvage or potentially curative treatment or the 7 who had bilateral recurrences (3 underwent bilaterial neck dissection), all died during follow-up. One patient with ipsilateral neck metastasis undergoing supraomohyoid dissection and presenting with tumor recurrence was unavailable for follow-up 8 months after such surgery. Eleven (25.0%) of the 44 patients with ipsilateral recurrences and 16 (51.6%) of the 31 with contralateral recurrences underwent classic or modified radical neck dissection. At study end, only 2 with ipsilateral and 3 with contralateral metastases were alive and without evidence of tumor recurrence.

A significant difference in survival was found according to the time from initial treatment to detection of neck treatment failure; 5-year SARs were 22.5% for the patients with an interval of 6 months or less, and 0% for those with an interval of more than 24 months (P<.001). Patients previously undergoing treatment of the neck exhibited poorer SARs than those not previously treated. The survival after neck recurrence was influenced by the type of previous neck dissection (P<.001) and previous postoperative radiotherapy (P = .003) (Table 2). Patients undergoing salvage radical classic or modified neck dissection exhibited higher SAR than did those treated by means of other techniques or those untreated. There was no significant impact of radiotherapy on the prognosis of patients with neck recurrences (Table 3).

Local and neck recurrences are the most frequent, and potentially preventable, sites of treatment failures in patients with oral carcinomas.^4,11-13 Our study confirms that most neck recurrences (in treated necks) or neck metastasis in patients not initially undergoing neck dissections are diagnosed within 2 years after initial surgical treatment.^13-15

In this experience, only 36 (44.0%) of the patients with neck recurrences were candidates for salvage surgical treatment with curative intent. Patients previously undergoing neck dissections underwent wide surgical resection of recurrent neck tumors, and none survived 5 years. A radical (classic or modified) neck dissection was the treatment of choice for patients with recurrences in a previously untreated side of the neck.

More than one third of our patients (37.8%) could not be treated owing to the extension of neck recurrence, poor performance status, or concomitant diagnosis of distant metastasis. A total of 18.3% of the patients underwent radiotherapy with palliative or curative intent, because of recurrence extension (unresectable tumors) or poor clinical status (untreatable patients). The high number of patients not eligible for salvage treatment of neck recurrences has been reported by others.^5,13,16-18 This constitutes the main argument in favor of elective neck dissections, particularly when follow-up difficulties can be anticipated.^5,19 However, several others reported higher rates of resectability and survival results similar to those of patients undergoing elective treatment of the neck initially.^4,13,20,21 These differences can be explained by follow-up difficulties, especially in socially disadvantaged patients with head and neck cancer who live far from the referral head and neck surgical center. These results suggest that patients at a high risk for neck metastasis should undergo an elective neck dissection, and patients with low-risk necks (T1 tumors with thickness of <3 mm) should undergo image investigation to confirm the neck status and then very close follow-up.

Our results confirmed that patients with neck recurrences have a poor prognosis, despite salvage treatment. The survival results in patients with recurrences diagnosed within 6 months in a previously untreated side of the neck were far superior to those observed in patients with recurrences diagnosed after a longer time. Patients previously undergoing neck dissection and with recurrences diagnosed after 6 months were not usually candidates for curative salvage treatment and were under a high risk for death. Only 5 patients with recurrences in a previously untreated side of the neck in the present study survived more than 5 years after salvage treatment. Important difficulties deter performance of a reliable clinical examination of the neck in patients previously undergoing operation or radiotherapy. The same problems have been noted by others.^13,15 Most neck recurrences are diagnosed at an advanced stage, with fixation to deep-neck structures and carotid artery. However, because of noticeable developments in surgical technique (resection and reconstruction of the carotid artery) and radiotherapy (mainly brachytherapy), some patients with advanced disease can now undergo salvage treatment, yet the survival results are disappointing.¹⁵ In the present study, a few patients survived 5 years after neck recurrences, and none survived 10 years. Cady and Catlin²² had only 1 surviving patient among 41 patients with epidermoid carcinoma of the gum undergoing a salvage neck treatment. Others using surgery or radiotherapy^23-25 also presented disappointing results.

A modified or classic radical neck dissection in patients with recurrences in a previously untreated side of the neck offered the best chance for survival (36.0% 5-year survival for patients with ipsilateral metastasis, and 30.8% for patients with contralateral metastasis). These results are similar to those registered in patients with clinically positive neck metastasis initially undergoing a neck dissection. However, they only express the survival of the patients who were candidates for surgical salvage treatment. The major problem is that a significant number of patients were not eligible for such treatment and died in a short time, and the survival results, including all patients, were significantly lower than those expected to be obtained with elective treatment of a cN0 neck but with positive pathological findings (pN+). I agree with Grandi et al¹³ and DeSanto and Beahrs²⁵ that the accuracy in the performance of the neck dissection at the initial treatment is very important, because recurrences in a previously treated side of the neck are difficult to detect and treat, and prognosis is extremely poor.

Accepted for publication August 24, 2001.

Corresponding author and reprints: Luiz P. Kowalski, MD, PhD, Head and Neck Surgery and Otorhinolaryngology Department, Centro de Tratamento e Pesquisa Hospital do Câncer A C Camargo, Rua Prof. Antonio Prudente, 211, 01509-010 São Paulo, Brazil (e-mail: lp_kowalski@uol.com.br).