Thirty-two patients were included in the study; all had stage IVb diseases (T4bN3 = 12 patients, T4bN2 = 9 patients, and T3 N3 = 11 patients). The median follow-up time was 18 months (range, 6–35.5 months). According to our unresectability criteria, there were 9 patients with prevertebral fascia invasion, 17 with carotid artery encasement, 5 with both prevertebral fascia invasion and carotid artery encasement, and 1 with mediastinal structure involvement. The patient’s clinical characteristics are presented in Table 1. After 3 cycles of induction chemotherapy, none of the patients achieved complete tumor response. However, 21 patients had partial tumor response, 7 had stable disease, and 4 had progressive disease.
Pathological examination of the surgical specimens revealed free, close, and positive margins in 1, 7, and 10 cases, respectively. Pathological extracapsular extension of the resected lymph node was observed in 9 cases.
Postoperative chemoradiotherapy was recommended in all surgical cases. Thirteen patients were able to complete the 6 cycles of chemotherapy, while 3 complete 1–3 cycles. Two patients had radiotherapy alone owing to a tumor-free surgical margin and no extracapsular extension of the dissected lymph node. Five patients with unresectable diseases were treated with concurrent chemoradiotherapy, 3 with radiotherapy, and 6 with supportive care. Pulmonary metastasis was noted in 3 patients during the follow-up period (12 months after surgery and adjuvant chemoradiotherapy in 1 patient, and 5 and 10 months after initiating supportive care in the other 2 patients, respectively).
Following induction chemotherapy, 18 patients with a partial response had sufficient tumor reduction and were considered resectable (i.e., the resectable group). The remaining 14 patients still had unresectable diseases (and comprised the unresectable group). The resectability rate was 56.3%.
Factors predicting tumor resectability
Clinical variables including primary site, T stage, N stage, the cause of unresectability, chemotherapy regimen, and tumor differentiation were analyzed to determine tumor resectability after induction chemotherapy. On univariate analysis, laryngeal cancer and TPF regimen were the factors associated with tumor resectability (p = 0.048, and p = 0.006, respectively). Multivariate analysis showed that receiving a TPF regimen was the only predictive factor associated with producing sufficient tumor reduction; patients who underwent TPF had a tumor resectability rate of 78.6%, while those who underwent PF and CP regimens had resectability rates of 40% and 37.5%, respectively (p = 0.044). (Table 2)When considering the characteristics of resectable tumors, those with carotid artery encasement had the highest chance of undergoing surgical resection, with a rate of 70.6%. However, tumors with prevertebral fascia invasion alone, prevertebral fascia invasion plus carotid artery encasement, and mediastinal structure involvement had resection rates of 44.4%, 40%, and 0%, respectively; the differences were not significant (p = 0.088).
Overall survival (OS)
The median OS of all patients was 16 months (range, 9.5–35.5 months). The median OS rates of the resectable and unresectable groups were 20.0 months (range, 16.0–35.5 months), and 9.5 months (range, 6.0–15.0 months), respectively (p = 0.008).
The estimated 2-year OS of all patients was 39.1% (95% confidence interval [CI], 22.1–55.7%). The estimated 2-year OS rates of the resectable and unresectable groups were 59.5% (95% CI, 33.2–78.3%), and 10.7% (95% CI, 1.1–35.4%), respectively (p = 0.0008) (Fig. 1).
Disease-free survival (DFS)
The median DFS of all patients was 13.5 (range, 7.5–21.5 months). The median DFS of the resectable and unresectable groups were 20.0 months (range, 12.5–25.0 months), and 6.0 months (range, 5.0–11.0 months), respectively (p = 0.009).
The estimated 2-year DFS of all patients was 36% (95% CI, 19.6–52.7%). The estimated 2-year DFS of the resectable and unresectable groups were 53.5% (95% CI, 27.9–73.6%), and 14.3% (95% CI, 2.3–36.6%), respectively (p = 0.0009) (Fig. 2).
Locoregional recurrence occurred in 20 patients; 8 (44.4%) were in the resectable group and 12 (85.7%) in the unresectable group.
Factors influencing OS and DFS
Clinical variables including age, sex, performance status, smoking, alcohol usage, primary site, T stage, N stage, the cause of unresectability, chemotherapy regimen, surgical resection, and tumor differentiation were analyzed to determine factors influencing OS and DFS rates in all patients. Univariate analysis revealed factors positively affecting OS and DFS rates were patients with performance status = 0 (p = 0.041, and p = 0.038, respectively), laryngeal cancer (p = 0.009, and p = 0.012, respectively), receiving the TPF regimen (p = 0.006, and p = 0.009, respectively), and surgical resection (p = 0.010, and p = 0.008, respectively).
However, on multivariate analysis, factors positively affecting the OS and DFS rates in all patients were surgical resection (p = 0.007 and 0.007, respectively), laryngeal primary (p = 0.009 and 0.005, respectively), and receiving the TPF regimen (p < 0.001 and = 0.006, respectively). (Table 3 and Table 4) No factor influencing OS and DFS were identified in the resectable group, while definitive chemoradiotherapy extended DFS in the unresectable group (p = 0.001).