Salvage therapy in relapsed squamous cell carcinoma of the oral cavity: How and when?
Abstract
BACKGROUND.
Relapse of tumors in patients with oral cavity squamous cell carcinoma (OSCC) is associated with a poor prognosis. In addition, salvage therapy may be a significant source of morbidity in patients with relapsing OSCC. The objective of the current study was to determine prognostic factors that predict which patients may benefit from such treatment.
METHODS.
From 953 patients who underwent primary radical surgery between 1996 and 2005, 272 patients with early-relapsed OSCC (n = 161) or late-relapsed OSCC (n = 111) were identified. The optimum cutoff point for relapse was chosen on the basis of 5-year disease-specific survival (DSS) and overall survival (OS).
RESULTS.
The optimal cutoff value for relapse was 10 months. Late relapses were associated with a better prognosis than relapses that occurred within the first 10 months (P < .0001 for both 5-year DSS and 5-year OS). Among patients with early-relapsed OSCC, a primary tumor depth <10 mm was associated significantly and independently with a better 5-year DSS (P = .014) and OS (P = .011). Among patients with late-relapsed OSCC, neck recurrence was a significant risk factor for adverse outcomes (P < .001 for both 5-year DSS and 5-year OS).
CONCLUSIONS.
A late relapse was associated with better survival than a relapse that occurred within the first 10 months. Patients with late-relapsed OSCC may benefit from salvage therapy, especially those who have a local recurrence. Among patients with early-relapsed OSCC, salvage therapy should be considered for those who have a primary tumor depth <10 mm. Cancer 2008. © 2007 American Cancer Society.
Survival in patients with oral cavity squamous cell carcinoma (OSCC) depends heavily on locoregional control. Although the overall neck and local control rates at 5 years are as high as 85%, the prognosis worsens precipitously in the presence of relapse or distant metastases.1-5 Once a relapse occurs, salvage therapy may prove valuable. However, salvage regimens may be a significant source of morbidity in patients with relapsing OSCC.
We demonstrated previously that a late relapse is associated with a better outcome than a relapse that occurs within the first months from the time of initial treatment in patients with tumors classified as 4b (T4b) OSCC.6 Nonetheless, some patients with late-relapsed OSCC may have an unfavorable prognosis. Thus, it is posited that the underlying tumor biology may play a role in clinical outcomes for patients with OSCC.
In view of the significant morbidity of salvage therapy,2, 3 it is important to identify the subset of OSCC patients who are ideal candidates for salvage regimens. The objectives of our current study were 1) to investigate survival for patients with early- and late-relapsed OSCC and identify the optimum cutoff point for relapse, 2) to determine prognostic factors that predict which patients with early-relapsed OSCC may benefit from salvage therapy, and 3) to investigate whether a subset of patients with late-relapsed OSCC may have a worse prognosis after salvage therapy.
MATERIALS AND METHODS
Patients
We retrospectively reviewed 955 consecutive patients with OSCC who underwent primary radical surgery between January 1996 and April 2005. Patients who had positive surgical margins (n = 2 patients) were excluded. Of the 953 study participants, 272 patients (28.5%) with first-relapsed OSCC were identified. All patients provided their informed consent to participate in the study. Patients in this series underwent an extensive presurgical evaluation. This evaluation included a medical history and complete physical examination, flexible fiberoptic pharyngoscopy, complete blood count and routine blood biochemistry, computed tomography scans or magnetic resonance images of the head and neck, chest radiographs, bone scan, and liver ultrasound. Staging was assigned according to the 1997 American Joint Committee on Cancer (AJCC), 5th edition staging criteria.7 There were 2 main reasons for the use of AJCC 1997 criteria instead of the AJCC 2002 staging system. First, some histopathologic specimens that were collected before 2002 were not available for further review. Second, correct staging of pathologic T4b (pT4b) disease according to the AJCC 2002 criteria8 may prove troublesome for retrospective analyses.5, 6, 9 In this study, staging was assigned according to pathologic status at primary treatment. The primary definitive treatment was performed as described previously.4-6, 9
Determination of the Optimal Cutoff Point for Early Versus Late Relapse
The 5-year disease-specific (DSS) and overall survival (OS) rates were calculated by using Kaplan-Meier analysis between 3 months and 24 months after primary definitive treatment. The optimum cutoff point for relapse was chosen on the basis of 5-year DSS and OS.
Salvage Therapy for Locoregional Recurrence
In the current study, 212 patients were considered salvageable (119 patients with early-relapsed OSCC and 93 patients with late-relapsed OSCC). One hundred thirty-four patients (63.2%) underwent salvage therapy (65 early-relapsed patients and 69 late-relapsed patients). Because of their unwillingness to undergo salvage therapy, 78 study participants (36.8%) had palliation therapy (54 early-relapsed patients and 24 late-relapsed patients). Patients with local tumor recurrence underwent radical surgical excision with safety margins ≥1.5 cm (all peripheral and deep margins). Patients who had a cervical lymph node recurrence underwent complete neck dissection. Lymph node groups from Level I through Level V were included in these dissections. Patients received postoperative adjuvant radiation therapy (RT), chemotherapy, or concurrent chemoradiotherapy (CCRT) when necessary on the basis of consensus reached by a Tumor Board conference. The treatment principle for patients with initial failure at distant sites was palliation, and the only exception was the presence of a single resectable site in the lung.
Data Analysis
Follow-up was continued until April 2007. All patients received follow-up examinations for at least 24 months after surgical treatment or until death. Descriptive statistics were summarized using frequencies, percentages, medians, standard deviations, and ranges. The Kaplan-Meier method was applied for survival analysis. The statistical significance was evaluated using the log-rank test. Univariate analyses (UVA) and multivariate analyses (MVA) were used to define independent risk factors. MVA of the prognostic factors were performed using the Cox logistic regression method with forward selection. All P values were 2-sided, and P values <.05 were considered significant.
RESULTS
Patients
The median age of the 272 study participants was 49 years (range, 25–82 years). One hundred thirty-three patients had local recurrences at a median of 15 months (range, 2–107 months), 139 patients had neck recurrences at a median of 7 months (range, 1–67 months), and 87 patients developed distant metastases at a median of 8 months (range, 1–75 months) after initial treatment. In total, 246 patients (90.4%) underwent neck dissection, including 206 ipsilateral dissections and 40 bilateral dissections. In total, 178 patients (65.4%) received postoperative adjuvant therapy with RT or CCRT.
Optimal Cutoff Value for Early Versus Late Relapse
Table 1 summarizes the 5-year DSS and OS rates across different cutoff points for tumor recurrence (expressed as months after initial treatment) in the 272 study participants. The cutoff points in the range between 4 months and 15 months yielded significant results (all P < .001). The cutoff point of 10 months was chosen, because it yielded the worst 5-year DSS and OS rates. According to this cutoff value, 161 patients had early relapses, and 111 patients had late relapses.
| Time to relapse after initial treatment, mo | 5-year DSS | 5-year OS | ||||
|---|---|---|---|---|---|---|
| ≤ (%) | > (%) | P | ≤ (%) | > (%) | P | |
| 4 | 22.7 | 32.7 | <.0001 | 19.6 | 31.9 | <.0001 |
| 5 | 20.1 | 34.7 | <.0001 | 18.1 | 34 | <.0001 |
| 6 | 18.7 | 37.4 | <.0001 | 17.2 | 36.6 | <.0001 |
| 7 | 16.8 | 41.5 | <.0001 | 15.5 | 40.8 | <.0001 |
| 8 | 14.5 | 47.2 | <.0001 | 13.4 | 46.5 | <.0001 |
| 9 | 15.1 | 49 | <.0001 | 13.2 | 49.5 | <.0001 |
| 10 | 13.8 | 53.5 | <.0001 | 12.2 | 54 | <.0001 |
| 11 | 15 | 55.1 | <.0001 | 13.4 | 55.7 | <.0001 |
| 12 | 15.2 | 58.4 | <.0001 | 13.8 | 59.1 | <.0001 |
| 13 | 15.9 | 59.2 | <.0001 | 14.5 | 59.9 | <.0001 |
| 14 | 16.4 | 61.5 | <.0001 | 15 | 62.3 | <.0001 |
| 15 | 17.5 | 62.6 | <.0001 | 16.1 | 63.5 | <.0001 |
- DSS indicates disease-specific survival; OS, overall survival.
Table 2 displays the general characteristics of patients who had early relapses compared patients who had late relapses. Compared with patients who had late-relapsed OSCC, patients who had early-relapsed OSCC had a significantly younger age at onset (P = .010), and they had an increased prevalence of moderate/poor cell differentiation (P = .038), pathologic 2b lymph node (pN2b) status (P < .001), extracapsular spread (P < .001), pathologic stage IV disease (P = .011), neck recurrence (P < .001), and distant metastases (P < .001). In addition, there was a significantly higher incidence of local recurrence in the group of patients with late-relapsed OSCC. Thus, salvage therapy was performed more frequently in patients who a late relapse compared with patients who had an early relapse (P < .001). Figure 1 shows that a late relapse was associated with better survival than a relapse that occurred within the first 10 months after initial treatment (P < .0001 for both 5-year DSS and 5-year OS).
(Left) Five-year disease-specific survival in patients with squamous cell carcinoma of the oral cavity (OSCC) who had early and late relapses. (Right) Five-year overall survival in patients with OSCC who had early and late relapses.
| Characteristic | No. of patients | P | |
|---|---|---|---|
| Early relapse (n = 161) | Late relapse (n = 111) | ||
| Sex | .590 | ||
| Men | 153 | 107 | |
| Women | 8 | 4 | |
| Age, y | .010 | ||
| ≤40 | 39 | 13 | |
| >40 | 122 | 98 | |
| Betel quid chewing | .264 | ||
| No | 30 | 15 | |
| Yes | 131 | 96 | |
| Cigarette smoking | .160 | ||
| No | 18 | 19 | |
| Yes | 143 | 92 | |
| Alcohol drinking | .594 | ||
| No | 53 | 40 | |
| Yes | 108 | 71 | |
| Tumor subsites | .816 | ||
| Tongue | 68 | 39 | |
| Mouth floor | 5 | 4 | |
| Lip | 5 | 2 | |
| Buccal | 52 | 38 | |
| Gum | 15 | 16 | |
| Hard palate | 8 | 6 | |
| Retromolar | 8 | 6 | |
| Neck dissection | .499 | ||
| No | 17 | 9 | |
| Yes | 144 | 102 | |
| Differentiation | .038 | ||
| Well | 45 | 47 | |
| Moderate | 93 | 54 | |
| Poor | 23 | 10 | |
| Primary treatment | .069 | ||
| Surgery | 49 | 45 | |
| Surgery plus RT | 63 | 45 | |
| Surgery plus CCRT | 49 | 21 | |
| Pathologic tumor classification | .312 | ||
| T1 | 18 | 17 | |
| T2 | 55 | 41 | |
| T3 | 44 | 20 | |
| T4 | 44 | 33 | |
| Pathologic lymph node status | <.001 | ||
| pNx (no neck dissection) | 17 | 9 | |
| pN0 | 43 | 54 | |
| pN1 | 18 | 21 | |
| pN2a | 1 | 0 | |
| pN2b | 71 | 22 | |
| pN2c | 11 | 5 | |
| Extracapsular spread | <.001 | ||
| No | 86 | 82 | |
| Yes | 75 | 29 | |
| Pathologic stage | .011 | ||
| I | 16 | 12 | |
| II | 22 | 27 | |
| III | 22 | 24 | |
| IV | 101 | 48 | |
| Local recurrence | <.001 | ||
| No | 106 | 33 | |
| Yes | 55 | 78 | |
| Neck recurrence | <.001 | ||
| No | 57 | 76 | |
| Yes | 104 | 35 | |
| Distant metastases | <.001 | ||
| No | 98 | 87 | |
| Yes | 63 | 24 | |
| Salvage with surgery and/or RT/CCRT at any site of locoregional recurrence | <.001 | ||
| No | 96 | 42 | |
| Yes | 65 | 69 | |
- RT indicates radiotherapy; CCRT, concomitant chemoradiotherapy.
Clinical Course
Patients with early relapse
At the time of the current analyses, 21 of 161 patients (13%) who had early-relapsed OSCC remained alive, and 140 patients (87%) had died. Of 161 patients, 23 patients (14.3%) had local recurrence only, 54patients (33.5%) had neck recurrence only, 31 patients (19.3%) had distant metastases only, 21 patients (13%) had local and neck recurrences, 3 patients (1.9%) had local and distant metastases, 21 patients (13%) had neck and distant metastases, and 8 patients (5%) had locoregional and distant metastases. One hundred thirty patients (80.7%) had local, neck, or locoregional recurrence. Sixty-five patients (50%) underwent salvage therapy, 51 patients underwent salvage surgery either alone or with adjuvant RT/CCRT, and the remaining 14 patients were salvaged with RT/CCRT. Among the 65 patients who were salvaged, 21 patients remained alive, and 44 patients died. None of early-relapsed patients with distant metastases met the criteria for salvage therapy, and they all received palliation therapy.
Patients with late relapse
At the time of the current analyses, 56 of 111 patients (50.5%) with late-relapsed OSCC remained alive, and 55 patients (49.5%) had died. Of the 111 late-relapsed patients, 56 patients (50.5%) had local recurrence only, 15 patients (13.5%) had neck recurrence only, 16 patients (14.4%) patients had distant metastases only, 16 patients (14.4%) had both local and neck recurrences, 4 patients (3.6%) had local and distant metastases, 2 patients (1.8%) had neck and distant metastases, and 2 patients (1.8%) had locoregional and distant metastases. Ninety-five patients (85.6%) had local, neck, or locoregional recurrence. Sixty-nine patients (72.6%) underwent salvage therapy, 58 patients underwent salvage surgery either alone or with adjuvant RT/CCRT, and the remaining 11 patients were salvaged with RT/CCRT. Among the 69 patients who were salvaged, 52 patients remained alive, and 17 patients died. Among the 24 patients with OSCC who had distant metastasis, only 1 patient met the criteria for salvage therapy. Thus, the remaining 23 patients received palliation therapy.
Prognosticators for Early Versus Late Relapse
Early relapse
UVA of the prognostic factors associated with 5-year DSS and OS rates in 161 patients with early-relapsed OSCC was performed. Table 3 shows that 27 parameters were subjected to UVA. MVA identified the variables that significantly affected both 5-year DSS and OS rates before salvage therapy: the presence of pathologic stage III or IV disease, tumor depth ≥10 mm, poor differentiation, and the presence of distant metastases. Conversely, tumor depth ≥10 mm, the presence of distant metastases, and the absence of salvage therapy were associated independently with both 5-year DSS and OS rates after salvage regimen (Table 4).
| Risk factor (no. of patients) | 5-Year DSS | 5-Year OS | ||
|---|---|---|---|---|
| Rate, % | P | Rate, % | P | |
| Sex | ||||
| Men (153) | 14.4 | .4160 | 12.7 | .5040 |
| Women (8) | 0 | 0 | ||
| Age, y | .6705 | .4107 | ||
| ≤40 (39) | 19.2 | 19.2 | ||
| >40 (122) | 12.2 | 10.3 | ||
| Betel quid chewing | .6313 | .5821 | ||
| No (30) | 18.5 | 17.9 | ||
| Yes (131) | 12.7 | 11 | ||
| Cigarette smoking | .9178 | .7328 | ||
| No (18) | 13 | 13 | ||
| Yes (143) | 13.9 | 12.1 | ||
| Alcohol drinking | .3691 | .4021 | ||
| No (53) | 10.4 | 10 | ||
| Yes (108) | 15.4 | 13.2 | ||
| Neck dissection | .0017 | .0007 | ||
| No (17) | 36.4 | 36.4 | ||
| Yes (144) | 11.2 | 9.6 | ||
| Treatment | .0001 | <.0001 | ||
| Surgery (49) | 25.3 | 24.2 | ||
| Surgery plus RT/CCRT (112) | 8.5 | 7.1 | ||
| Postsurgery RT, d* | .2123 | .0919 | ||
| ≤42 (120) | 15.9 | 14.5 | ||
| >42 (41) | 6.7 | 5 | ||
| Pathologic tumor classification | .0002 | .0001 | ||
| T1–T2 (73) | 22.6 | 20.5 | ||
| T3–T4 (88) | 6.3 | 5.7 | ||
| Pathologic lymph node status* | <.0001 | <.0001 | ||
| pN0 (60) | 30.8 | 25.6 | ||
| pN1–pN2 (101) | 3 | 2.6 | ||
| Pathologic overall stage | <.0001 | <.0001 | ||
| I–II (38) | 40.7 | 37.2 | ||
| III–IV (123) | 5.2 | 4.7 | ||
| Differentiation | .0001 | .0001 | ||
| Well/moderate (138) | 15.3 | 13.6 | ||
| Poor (23) | 4.6 | 4.4 | ||
| Bone invasion | .6495 | .5211 | ||
| No (133) | 14.4 | 12.7 | ||
| Yes (28) | 11.9 | 10.7 | ||
| Skin invasion | .0071 | .0272 | ||
| No (149) | 15 | 13.2 | ||
| Yes (12) | 0 | 0 | ||
| Nerve invasion† | .2234 | .1970 | ||
| No (92) | 16.3 | 14.9 | ||
| Yes (68) | 10.7 | 8.8 | ||
| Vessel invasion† | .5431 | .6190 | ||
| No (154) | 14.5 | 12.8 | ||
| Yes (5) | 0 | 0 | ||
| Lymph invasion† | .9842 | .9600 | ||
| No (141) | 14.3 | 12.7 | ||
| Yes (18) | 12.5 | 11.1 | ||
| Close margin, mm† | .3610 | .5215 | ||
| >4 (132) | 18.2 | 12.6 | ||
| ≤4 (23) | 14.5 | 18.2 | ||
| Tumor depth, mm† | .0001 | <.0001 | ||
| ≤10 | 26.9 | 25.5 | ||
| >10 | 6.9 | 5.7 | ||
| Extracapsular spread | <.0001 | <.0001 | ||
| No (86) | 23.4 | 20.9 | ||
| Yes (75) | 2.9 | 2.7 | ||
| Level IV and/or V metastases | .0373 | .0613 | ||
| No (145) | 14.8 | 13 | ||
| Yes (16) | 6.3 | 6.3 | ||
| Pathologic N2c lymph node status | .2622 | .1791 | ||
| No (150) | 14.7 | 13.2 | ||
| Yes (11) | 0 | 0 | ||
| Local recurrence | .9527 | .9712 | ||
| No (106) | 17.3 | 15 | ||
| Yes (55) | 7.8 | 7.3 | ||
| Neck recurrence | .0020 | .0078 | ||
| No (57) | 7.2 | 7 | ||
| Yes (104) | 17.7 | 15.2 | ||
| Distant metastases | <.0001 | <.0001 | ||
| No (98) | 22.5 | 19.4 | ||
| Yes (63) | 0 | 0 | ||
| Second primary tumor | .1692 | |||
| No (149) | 11.2 | |||
| Yes (12) | 25 | |||
| Salvage with surgery and/or RT/CCRT at any locoregionalrecurrence | <.0001 | <.0001 | ||
| Yes (65) | 38.4 | 35.8 | ||
| No (96) | 0 | 0 | ||
- DSS indicates disease-specific survival; OS, overall survival; RT, radiotherapy; CCRT, concurrent chemoradiotherapy.
- * Expressed as the number of patients who did received RT within 42 days. Patients with pNx disease were given a pN0 disease classification.
- † One patient had unknown nerve invasion, 2 patients had unknown vessel invasion, 2 patients had unknown lymph invasion, 6 patients had unknown margins, and 1 patient had unknown tumor depth.
| Risk Factor | Before salvage | After salvage | ||||||
|---|---|---|---|---|---|---|---|---|
| 5-Year DSS | 5-Year OS | 5-Year DSS | 5-Yearr OS | |||||
| P | HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | |
| Pathologic stage III–IV | .004 | 1.500 (1.137–1.978) | .003 | 1.492 (1.144–1.946) | NS | NS | ||
| Poor differentiation | .025 | 1.026 (1.003–1.049) | .016 | 1.027 (1.005–1.049) | NS | NS | ||
| Tumor depth ≥10 mm | .023 | 1.051 (1.007–1.097) | .016 | 1.053 (1.010–1.098) | .014 | 1.052 (1.010–1.095) | .011 | 1.053 (1.012–1.095) |
| Distant metastases | <.001 | 1.257 (1.141–1.384) | <.001 | 1.211 (1.103–1.330) | <.001 | 1.270 (1.158–1.393) | <.001 | 1.220 (1.115–1.335) |
| No salvage therapy | NA | NA | <.001 | 3.325 (2.094–5.280) | <.001 | 3.379 (2.155–5.298) | ||
- DSS indicates disease-specific survival; OS, overall survival; HR, hazard ratio; 95% CI, 95% confidence interval; NS, not significant; NA, not applicable.
Late relapse
UVA of the prognostic factors associated with 5-year DSS and OS rates in 111 patients with late-relapsed OSCC was performed. Table 5 shows that 27 parameters were subjected to UVA. MVA identified the variables that significantly affected both 5-year DSS and OS rates before salvage therapy: poor differentiation, the presence of distant metastases, pT3 or pT4 disease, extracapsular spread, and neck recurrence. Conversely, the presence of distant metastases, the absence of salvage therapy, and neck recurrence were associated independently with both 5-year DSS and OS rates after salvage regimen (Table 6).
| Risk factor (no. of patients) | 5-Year DSS | 5-Year OS | ||
|---|---|---|---|---|
| Rate, % | P | Rate, % | P | |
| Sex | .9346 | .8618 | ||
| Men (107) | 53.6 | 54.2 | ||
| Women (4) | 50 | 50 | ||
| Age, y | .0951 | .0563 | ||
| ≤40 (13) | 46.2 | 46.2 | ||
| >40 (98) | 54.6 | 55.2 | ||
| Betel quid chewing | .7383 | .6215 | ||
| No (15) | 60 | 60 | ||
| Yes (96) | 52.3 | 53 | ||
| Cigarette smoking | .7216 | .7867 | ||
| No (19) | 48.1 | 48.1 | ||
| Yes (92) | 54.6 | 55.2 | ||
| Alcohol drinking | .5482 | .6839 | ||
| No (40) | 55.9 | 57.4 | ||
| Yes (71) | 52.2 | 52.2 | ||
| Neck dissection | .5230 | .4651 | ||
| No (9) | 64.8 | 64.8 | ||
| Yes (102) | 52.7 | 53.2 | ||
| Treatment | .0055 | .0031 | ||
| Surgery (45) | 69.8 | 68.6 | ||
| Surgery plus RT/CCRT (66) | 41.9 | 43.9 | ||
| Postsurgery RT, d* | .0140 | .0174 | ||
| ≤42 (91) | 59.7 | 60.5 | ||
| >42 (20) | 27.8 | 27.8 | ||
| Pathologic tumor classification | .0003 | .0013 | ||
| T1–T2 (58) | 69.5 | 68.9 | ||
| T3–T4 (53) | 35.7 | 37.7 | ||
| Pathologic lymph node status* | .0003 | .0004 | ||
| pN0 (63) | 67.1 | 66.6 | ||
| pN1–N2 (48) | 34.8 | 38 | ||
| Pathologic overall stage | .0001 | .0002 | ||
| I–II (39) | 78.2 | 77.6 | ||
| III–IV (72) | 40 | 41.5 | ||
| Differentiation | .0392 | .0514 | ||
| Well/moderate (101) | 57.3 | 58 | ||
| Poor (10) | 20 | 20 | ||
| Bone invasion | .2995 | .3733 | ||
| No (90) | 56.7 | 57.4 | ||
| Yes (21) | 39.4 | 39.4 | ||
| Skin invasion | .0436 | .0469 | ||
| No (101) | 55.5 | 56.1 | ||
| Yes (10) | 29.2 | 29.2 | ||
| Nerve invasion | .0346 | .0110 | ||
| No (81) | 60 | 59.2 | ||
| Yes (30) | 34.9 | 40 | ||
| Vessel invasion | .4520 | .4566 | ||
| No (110) | 53.1 | 53.7 | ||
| Yes (1) | 100 | 100 | ||
| Lymph invasion | .0061 | .0053 | ||
| No (101) | 57 | 58.2 | ||
| Yes (10) | 0 | 0 | ||
| Close margins, mm† | .2369 | .4153 | ||
| >4 (96) | 59.8 | 57.9 | ||
| ≤4 (13) | 36.9 | 36.9 | ||
| Tumor depth, mm | .0056 | .0019 | ||
| ≤10 (40) | 76.3 | 75.7 | ||
| >10 (71) | 43.2 | 41.9 | ||
| Extracapsular spread | <.0001 | <.0001 | ||
| No (82) | 69 | 66.1 | ||
| Yes (29) | 16.2 | 20.3 | ||
| Level IV and/or V metastases | .3174 | .3762 | ||
| No (110) | 56.1 | 54.6 | ||
| Yes (1) | 0 | 0 | ||
| Pathologic N2c | .3327 | .3297 | ||
| No (106) | 54.3 | 54.8 | ||
| Yes (5) | 40 | 40 | ||
| Local recurrence | .0001 | <.0001 | ||
| No (33) | 30.6 | 28.6 | ||
| Yes (78) | 65.9 | 65.1 | ||
| Neck recurrence | .0028 | .0063 | ||
| No (76) | 66.1 | 65 | ||
| Yes (35) | 32.2 | 30.1 | ||
| Distant metastases | <.0001 | <.0001 | ||
| No (87) | 66.2 | 65.8 | ||
| Yes (24) | 7.1 | 10.4 | ||
| Secondary primary tumors | .8555 | |||
| No (99) | 53.6 | |||
| Yes (12) | 58.3 | |||
| Salvage with surgery and/or RT, CCRT at any locoregional recurrence | <.0001 | <.0001 | ||
| Yes (69) | 80 | 79.2 | ||
| No (42) | 9.3 | 12.4 | ||
- DSS indicates disease-specific survival; OS, overall survival; RT, radiotherapy; CCRT, concurrent chemoradiotherapy.
- * Expressed as the number of patients who did received RT within 42 days. Patients who had pNx lymph node status were classified with pN0 disease.
- † Two patients had unknown margins.
| Risk factor | Before salvage | After salvage | ||||||
|---|---|---|---|---|---|---|---|---|
| 5-Year DSS | 5-Year OS | 5-Year DSS | 5-yr OS | |||||
| P | HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | |
| Poor differentiation | .003 | 1.056 (1.018–1.095) | .004 | 1.055 (1.018–1.095) | NS | NS | ||
| Extracapsular spread | .008 | 2.432 (1.264–4.677) | .003 | 2.657 (1.390–5.076) | NS | NS | ||
| pT3–pT4 | .002 | 1.610 (1.197–2.167) | .006 | 1.492 (1.120–1.986) | NS | NS | ||
| Distant metastases | <.001 | 1.386 (1.171–1.641) | <.001 | 1.348 (1.141–1.591) | <.001 | 1.435 (1.225–1.681) | <.001 | 1.397 (1.197–1.629) |
| Neck recurrence | .001 | 1.267 (1.095–1.465) | .003 | 1.239 (1.075–1.428) | <.001 | 1.315 (1.132–1.529) | <.001 | 8.536 (4.251–17.139) |
| No salvage therapy | NA | NA | <.001 | 8.398 (4.341–16.249) | <.001 | 7.662 (4.046–14.510) | ||
- DSS indicates disease-specific survival; OS, overall survival; HR, hazard ratio; 95% CI, 95% confidence interval; NA, not applicable; pT, pathologic tumor classification; NS, not significant.
Outcomes of Different Treatment Modalities After Relapse
The results of 3 different treatment modalities in patients with relapsed OSCC were analyzed. Patients were treated by salvage surgery (with or without RT/CCRT), salvage RT/CCRT, and palliative therapy. In patients with early-relapsed OSCC, a significant benefit was demonstrated for salvage therapy (P < .0001) in terms of both 5-year DSS and OS. However, no significant difference was observed for salvage surgery (either with or without RT/CCRT) compared with salvage RT/CCRT (Fig. 2). Similarly, in patients with late-relapsed OSCC, a significant benefit was demonstrated for salvage therapy (P < .0001) in terms of both 5-year DSS and OS. It is noteworthy that the outcome for patients who were salvaged with surgery (either with or without RT/CCRT) was significantly better compared with the outcome for patients who were salvaged with RT/CCRT in terms of both 5-year DSS (P = .0001) and OS (P = .0004) (Fig. 3).
(Left) Five-year disease-specific survival according to treatment modality in patients with squamous cell carcinoma of the oral cavity (OSCC) who had an early relapse. (Right) Five-year overall survival according to treatment modality in patients with OSCC who had an early relapse. RT indicates radiotherapy; CCRT, concurrent chemoradiotherapy.
(Left) Five-year disease-specific survival according to treatment modality in patients with squamous cell carcinoma of the oral cavity (OSCC) who had a late relapse. (Right) Five-year overall survival according to treatment modality in patients with OSCC who had a late relapse. RT indicates radiotherapy; CCRT, concurrent chemoradiotherapy.
DISCUSSION
Tumor relapse is not uncommon after primary treatment for OSCC. Salvage therapy may represent a valuable chance for survival in patients with relapsed OSCC. However, the outcome after salvage therapy remains unclear in this patient group, and it is difficult to determine whether the patients with locoregional recurrence, who are most likely to have distant metastases, may be salvaged. In addition, no data currently are available on the effectiveness of different salvage regimens, including salvage surgery and RT/CCRT, in patients with relapsed OSCC. The current study provides novel, important insights into these important issues.
First, we have demonstrated that patients with early-relapsed OSCC may benefit from salvage therapy in terms of both 5-year DSS and OS. However, no significant difference was observed for salvage surgery compared with salvage RT/CCRT, regardless of whether patients underwent local or neck recurrence. In addition, it also is noteworthy that our patients with early-relapsed OSCC had a higher prevalence of pN2b disease, pathologic stage IV tumors, extracapsular spread, neck recurrence, and distant metastases. Accordingly, early-onset relapse is associated with biologically more aggressive tumor behavior. Hence, salvage therapy may be of significant benefit in this patient group, regardless of which salvage regimen is selected. Conversely, the outcome of patients with late-relapsed OSCC who were salvaged with surgery was significantly better compared with the outcome of patients who were salvaged with RT/CCRT. In addition, outcomes were significantly better in patients who had a recurrence at the primary site compared with patients who had a recurrence in the neck. These findings in late-relapsed OSCC patients are in keeping with previous observations.1-3, 10, 11
Second, we have demonstrated that salvage therapy (either surgery or RT/CCRT) is associated with a good prognosis in patients who have an early relapse (Fig. 2). Moreover, salvage therapy with surgery is associated with a good prognosis in patients who have a late relapse (Fig. 3). Salvage surgery, however, leads to more technical difficulties compared with resection of the primary disease. In addition, an extensive resection with adequate safety margins is more important in the presence of relapsing tumors than in primary lesions. Reliable data concerning the treatment of recurrent OSCC are important, because the clinician frequently is confronted with the dilemma of whether to perform salvage therapy, with its associated morbidity. Our current results may permit clinicians to offer patients an informed choice before they make the decision to undergo salvage therapy for recurrent OSCC, although further prospective studies are needed to validate our findings.
Third, neck recurrence was identified as an independent, adverse prognostic factor in patients with late-relapsed OSCC, but not in patients with an early relapse. It also should be noted that salvage therapy may have inherent difficulties when tumor recurrence arises in the field of the previous neck dissection. A noteworthy observation from our study is that early-relapsed patients with neck recurrence had a better prognosis compared with patients who did not have a neck recurrence. In the current study, distant metastases occurred less frequently (P < .001) in early-relapsed patients who had a neck recurrence (27.9%; 29 of 104 patients) than in patients who did not have a neck recurrence (59.6%; 34 of 57 patients). Therefore, salvage therapy was used more frequently (P = .001) in the former patient group (83.7%; 41 of 49 patients) that in the latter group (28.6%; 2 of patients 7). We also demonstrated that late-relapsed patients with a local recurrence had better survival compared with patients in the same group who did not have a local recurrence. Similarly, patients who had late-relapsed OSCC with a local recurrence had a lower distant metastasis rate (P < .001;7.7%, 6 of 78 patients) compared with patients in the same group who did not have a local recurrence (54.5%; 18 of 33 patients). Thus, we posited that this patient group can undergo successful salvage therapy.
An important finding of our study is that patients with early-relapsed OSCC had a worse outcome than their counterparts with late-relapsed OSCC. Therefore, it is crucial to define a reliable cutoff point to distinguish early versus late relapse. It is also of paramount importance to establish whether different risk factor profiles may predict prognosis in patients who have an early relapse compared with patients who have a late relapse. Previous studies did not specifically analyzed recurrences as early as 6 months from the time of initial treatment.1, 2, 12 However, relapse within the first 6 months after primary treatment is not uncommon in patients with OSCC. Moreover, in the current study we chose to include all patients who underwent radical surgery and received adjuvant RT/CCRT, even in the presence of close margins (≤4 mm). Because the patients in this group generally have a poor prognosis, they should be considered as candidates to receive salvage therapy. In the light of our findings, the optimum cutoff value to discern between early versus late relapse was 10 months. Among patients with early-relapsed OSCC, salvage therapy should be considered those who have a primary tumor depth <10 mm. These findings are in line with previous reports from our group and others showing that tumor depth is a significant prognosticator for both local tumor control and survival rates.13-18 Moreover, we have provided evidence indicated that tumor depth is more important in OSCC patients who have an early relapse (P value significant in MVA) compared with a late relapse (P value significant in UVA but not in MVA) (Tables 4, 6).
Given the relatively high incidence of second primary cancers in the oral cavity in Taiwan,19 we did not exclude patients with these tumors from our study. Moreover, our current and previous data did not indicate any significant difference in survival rates between OSCC patients who had a salvageable second primary tumor in the oral cavity compared with OSCC patients who were without any second primary cancer.19
In conclusion, a late relapse was associated with better survival than a relapse that occurred within the first 10 months after initial treatment. Salvage surgery, either alone or in combination with adjuvant therapy, may be valuable for late-relapsed patients, especially in the presence of local recurrence. Salvage surgery and RT/CCRT had similar efficacy in patients with early-relapsed OSCC. In this patient group, salvage therapy should be considered for individuals who have a primary tumor depth <10 mm.
