Adding concurrent chemotherapy to postoperative radiotherapy improves locoregional control but Not overall survival in patients with salivary gland adenoid cystic carcinoma—a propensity score matched study

Despite being the most common form of salivary gland cancer, salivary gland adenoid cystic carcinoma (SGACC) is a rare malignancy which accounts for only 1 % of all head and neck tumors [1, 2]. Because of its local invasiveness, post-operative radiotherapy (PORT) is frequently used to increase locoregional control (LRC), particularly in patients bearing adverse prognostic factors, including positive surgical margins, advanced T stage, and perineural invasion (PNI) [3‐5]. Despite important advances in combination therapies, the 10 year rate of locoregional recurrence continues to remain high (~30 %) [3‐5]. Several multicenter randomized trials have demonstrated the effectiveness of post-operative chemoradiotherapy (POCRT) in patients with head and neck squamous cell carcinoma (HNSCC) [6‐8]. However, as a result of the rarity of SGACC, data concerning the role of adjuvant therapies in this clinical entity remain scarce. Although chemotherapy is generally reserved for the treatment of recurrent or metastatic SGACC [9], the clinical usefulness of adding concurrent chemotherapy to PORT remains unclear.

Based on the evidence supporting the clinical utility of POCRT in HNSCC patients [6‐8], we have been utilizing POCRT in our SGACC patients bearing adverse prognostic factors. In the present study, we retrospectively investigated the long-term outcomes of SGACC patients who were treated with surgery followed by PORT or POCRT in our institutions. Treatment outcomes between the PORT and POCRT groups were compared using propensity score matching in order to minimize biases. We also performed subgroup analyses with the goal of identifying specific SGACC patients who can benefit most from POCRT.

Despite its generally indolent course, locoregional recurrence and distant metastases remain major clinical issues in SGACC patients who bear adverse prognostic factors even when aggressive treatment modalities are used [3‐5, 13]. In patients with locally advanced HNSCC, POCRT has been shown to confer survival benefits and has been incorporated into the current standards of care [6, 7]. However, the question as to whether the addition of concurrent chemotherapy may improve outcomes in SGACC patients undergoing PORT remains open. Accordingly, the available data on the clinical utility of POCRT in patients with salivary gland cancers remain scarce and controversial. In a matched case–control study of salivary gland cancer patients who were treated either with platinum-based POCRT (N = 12) or PORT alone (N = 12), a significantly higher 3 year OS was observed in the POCRT arm (83 % vs 44 %) after a short follow-up period (median: 14.9 months for PORT and 31.6 months for POCRT) [14]. In another retrospective study of 35 patients undergoing adjuvant IMRT either with (N = 22) or without (N = 13) concurrent chemotherapy, no clear superiority for concurrent chemoradiation was observed after a median follow-up time of 2.3 years [15]. Because these studies included a markedly high number of different histological subtypes with small simple sizes and short follow-up periods, the question as to whether their findings can be generalizable to SGACC remains open. Because of the slowly progressive nature of SGACC, the median time to recurrence following adjuvant treatment is approximately 2–3 years [4, 5] making long-term follow-up particularly important to obtain reliable data.

In this study, we described the long-term outcomes of a multicenter series of SGACC patients treated with either POCRT or PORT. Besides performing a central review of all pathological specimens, we used propensity score matching to minimize the potential bias associated with a retrospective analysis. The results of our study demonstrate that SGACC patients treated with POCRT have significantly better LRC rates than those treated with PORT (5- and 8 year; 97 and 97 % versus 79 and 67 %, P = .017, Fig. 1b), particularly in presence of positive resection margins, PNI, or stage III − IV disease (Table 3). These findings indicate that concomitant chemotherapy has the potential to produce substantial radiosensitization of SGACC. Only a limited number of published studies have investigated the feasibility of chemoradiotherapy in SGACC patients, with the majority of them having been focused on the potential usefulness of definitive chemoradiotherapy. A case series of five non-resected SGACC patients managed with definitive carboplatin and paclitaxel-based chemoradiotherapy showed no locoregional recurrences after a median follow-up of 36 months, although one patient developed distant metastases at 7 months [16]. In a retrospective study that included 16 patients with locally advanced SGACC, the use of definitive chemoradiotherapy (with either intra-arterial or intravenous cisplatin or carboplatin) was associated with a 5 year local progression-free survival rate of 61 % after a median follow-up of 61 months [17]. In this report, distant metastasis was the most common pattern of treatment failure (five patients), and the 5 year OS and progression-free survival rates were 87 and 39 %, respectively. Taken together, these data suggest that platinum-based chemoradiation can exert a significant local cancer killing effect in SGACC.

In accordance with the published literature [3‐5, 17], distant metastasis was the main pattern of treatment failure and the most common cause of disease-specific mortality in our study. Although our findings indicate that the addition of concurrent chemotherapy to PORT significantly reduced locoregional relapses, we did not observe a decrease of hematogenous spread among these patients. Patients bearing multiple adverse prognostic factors were still at high risk of distant failure even with controlled locoregional disease. In addition, because of the slowly progressive course of SGACC, the median survival in patients with recurrence was reported to be as long as 2–3 years [18]. In our study cohort, the median survival time after a diagnosis of locoregional recurrence was 29 months (range: 7–108 months). These observations potentially explain the fact that the LRC benefit resulting from POCRT did not translate into a survival advantage.

Because the goals of comprehensive cancer care are not limited to maximize patient survival but also improve quality of life, we believe that the significant reduction in the rate of locoregional failure in SGACC patients should be considered clinically significant (especially because this type of failure is generally unsalvageable and symptomatic). Patients with locoregional relapses have several years of life expectancy but are particularly prone to symptoms that can seriously affect the overall quality of life (e.g., intractable pain). The time to first use of opioid analgesics is a meaningful clinical endpoint in the field of oncology and has been validated for assessing treatment efficacy in a number of previous prospective randomized trials [19‐22]. Intriguingly, propensity score analysis demonstrated that POCRT was significantly associated with a higher ORPFS (5 year: 96 % versus 82 %, P = .038, Fig. 2b). Subgroup analyses also demonstrated ORPFS improvement in patients with positive resection margins, PNI, or stage III − IV disease. These results suggest that an increase in LRC rates may be paralleled by a reduced risk of cancer-related pain (Table 3).

Despite every effort made to perform a very careful review of our data and multiple statistical examinations, there are still some unavoidable limitations inherent to our study. They include, but are not limited to, the non-randomized retrospective nature of the research and the long enrollment period. The relatively small number of events in this observational cohort and several unmeasured factors (including subjective treatment decisions and the varying surgical expertise over time) may have biased our results. In addition, we cannot exclude an underestimation of treatment-related side effects in patients who received concurrent chemotherapy in addition to postoperative radiotherapy. In this regard, several prospective randomized trials conducted in patients with head and neck malignancies have demonstrated that adjuvant chemoradiation can increase grade III − IV acute toxicities as compared with PORT alone (EORTC 22931: 41 % vs 21 %, P = 0.001; RTOG 9501: 77 % vs 34 %, P < 0.001) [6, 7] with an insignificant higher incidence of grade III − V late complications (RTOG 9501/intergroup long-term report: 24.9 % vs 20.5 %, P = 0.34) [23]. We are also aware that a median follow-up of 71 months is relatively short for the assessment of survival outcomes in patients with a slowly progressive malignancy like SGACC. However, its low prevalence makes prospective treatment trials difficult to conduct. The ongoing RTOG 1008 randomized trial has been designed to investigate the efficacy of cisplatin-based POCRT in patients with salivary gland carcinomas and a pathologic stage of T3-4, N1-3, or surgical margins ≤ 1 mm. However, this trial aims to enroll patients with a wide range of histological subtypes (including adenocarcinoma, mucoepidermoid carcinoma, salivary duct carcinoma, acinic cell carcinoma, and SGACC) until a sample size of 120 cases is achieved. Although the results of this trial will be paramount for clarifying the feasibility of POCRT in patients with salivary gland carcinomas, it is unclear whether the findings will be generalizable to SGACC. We therefore believe that our multicenter study using propensity score matching may add valuable information on the clinical usefulness of adding concurrent chemotherapy to PORT for SGACC patients.

The results from our study indicate that the addition of concurrent chemotherapy to PORT may increase LRC and ORPFS rates in SGACC patients, particularly in presence of stage III − IV disease, positive surgical margins, or PNI. Unfortunately, no statistically significant differences were observed in terms of DMFS, DFS, and OS. The occurrence of distant metastases remains a major clinical issue in patients bearing adverse prognostic factors. Consequently, further studies on the potential usefulness of concurrent chemotherapy and the clinical value of novel systemic agents for SGACC are urgently required.