Abstract
Background and purpose
Adenoid cystic carcinoma (ACC) is a rare salivary gland malignancy often characterized by an indolent course but significant risk of distant metastasis. The role of concurrent chemoradiotherapy (CRT) in improving oncologic outcomes remains controversial. This study aimed to assess the potential benefits of CRT compared with exclusive radiotherapy (RT) in ACC patients treated with curative intent.
Materials and methods
A retrospective cohort study was conducted using real-world data from a tertiary cancer center. Patients with head and neck ACC treated with curative RT between 2007 and 2022 were analyzed. Outcomes were evaluated using Kaplan–Meier and Cox regression analyses. Propensity score matching (PSM) was employed to control for confounding factors. The primary survival outcomes were distant metastasis-free survival (DMFS) and distant metastasis-free interval (DMFI).
Results
A total of 178 patients were included (89% receiving surgery), of whom 24 received CRT. Median follow-up was 85.2 months. In the unmatched cohort, CRT showed a trend toward improved DMFS (median 101.38 months versus 50.16 months, P = 0.052) and DMFI (101.38 months versus 53.25 months, P = 0.071). PSM analysis (n = 40) demonstrated statistically significant improvement in DMFI for CRT (median 101.38 months versus 39.8 months, P = 0.032; hazard ratio for distant metastasis 0.43, P = 0.037). No significant differences were observed in overall survival or locoregional control.
Conclusions
CRT may delay the onset of distant metastasis and extend DMFI in ACC patients, particularly younger and fit individuals. While overall survival benefits were not observed, these findings support CRT consideration in selected patients. Further prospective studies are warranted to confirm these results.
Key words: adenoid cystic carcinoma, chemoradiation, salivary gland cancer, chemotherapy, adjuvant radiotherapy, head and neck cancer
Highlights
-
•
Concurrent chemoradiotherapy may delay distant metastases in ACC patients.
-
•
This is the first study to explore DMFS and DMFI outcomes in ACC treated with CRT.
-
•
Propensity score matching revealed significant improvement in DMFI with CRT.
-
•
Real-world evidence supports CRT consideration for young, fit ACC patients.
Introduction
Salivary gland carcinomas (SGCs) are a wide group of epithelial malignancies arising from either major salivary glands or minor glands of the upper aerodigestive tract.1,2 These cancers are rare (incidence 1.135 × 100 000/year in Europe).1, 2, 3 Among the various SGC subtypes, adenoid cystic carcinoma (ACC) is the second most frequent SGC after mucoepidermoid carcinoma. Although having an indolent course, ACC patients can be affected by distant metastases even years after primary tumor diagnosis. The mainstay of the treatment of SGC in general, including ACC, is surgery. Radiotherapy is administered in the case of unresectable disease or after operation.4,5 Unlike squamous-cell carcinoma of the head and neck (HNSCC), for SGC patients, including ACC, there is no robust evidence of associating concomitant chemotherapy (CT) to photon-based intensity-modulated radiotherapy (IMRT). Given the difficulties in conducting randomized clinical trials in rare cancers and providing high-level evidence-based recommendations, the choice of delivering concomitant chemoradiation (CRT) instead of exclusive radiotherapy (RT) is usually established on a case-by-case basis after multidisciplinary discussion. This disease is even more difficult to study and treat because the literature indicates some variability in outcomes and the effectiveness of combined therapy.4 The available conflicting results underscore the need for further studies to better understand the specific role of CT in addition to RT in the curative treatment of ACC patients.
This study aimed to comprehensively assess whether a real-world database of a tertiary cancer center in Northern Italy, expert in the management of rare head and neck tumors including SGCs, could provide meaningful information about patient outcomes. Our institution is an international reference center for the management of rare head and neck cancers,6 and hundreds of ACC patients have been treated in the last decades.7 The exploitation of real-world data (RWD) that may be obtained retrospectively analyzing the experience of our tertiary cancer center may add further knowledge in the field of ACC natural history and exploring the role of concomitant CRT delivered in the curative setting.
Methods
Study objectives, design, data sources and variables
The research question of this work was to assess the contribution of CRT in the context of curative treatment of head and neck ACC patients. The target outcome was survival (details reported in Statistical analysis section).
The exposure of this analytical/explanatory research was the administration of concomitant CT in association with curative RT, delivered either as definitive treatment or after operation. The comparator group consisted of subjects receiving exclusive RT (i.e. without concomitant CT) with curative intent. CT consisted of platinum-based regimens, i.e. cisplatin-based (100 mg/m2 once every 21 days for up to 3 cycles or 40-50 mg/m2 once every 21 days for up to 6 cycles) or carboplatin-based [area under the curve (AUC) 5 once every 21 days for up to 3 cycles]. The study design used to address the research question was a retrospective cohort study.
Eligibility criteria used to select patients were access at our institution at any time of their natural history; confirmation of ACC diagnosis made by an expert pathologist; primary tumor located in the head and neck (i.e. ACCs originating from the trachea, lung, breast, skin, vulva, etc. were excluded); locoregional disease treated with curative intent; RT delivered in the context of curative treatments (i.e. either adjuvant or definitive); availability of full details about RT energy (i.e. photons, protons, carbon ions), technique (e.g. IMRT), dose and fractionation; availability of pretreatment clinical data and long-term oncologic outcomes (i.e. patients lost to follow-up were excluded).
The institutional data warehouse (DWH) of our institution was the original source of the RWD under study for this work. The DWH database is based on Oracle, and the electronic clinical record on SqlServer. Records updated in each single institutional clinical application during the day are uploaded to the DWH every night. Briefly, the DWH consists of the collection of all the institutional patient-oriented databases, with many coded variables with administrative aims, and fewer codes for clinical ones. The DWH includes also text-based variables collected for clinical aims from which clinical data can be extracted.
Details about the methodologies of data retrieval, DWH interrogation, natural language processing, data curation, and the clinical and pathological data collection have already been published,8 and are summarized in the Supplementary Material, available at https://doi.org/10.1016/j.esmorw.2025.100161.
Statistical analysis
Given the retrospective nature of the study, no preplanned sample size was calculated. Clinical data that could be obtained through the interrogation of the institutional DWH was scrutinized. Then, patients were selected following the study eligibility criteria (pathological diagnosis of ACC; locoregional disease treated with curative intent; RT delivered in the context of curative treatments (i.e. either adjuvant or definitive); availability of pretreatment clinical data; availability of post-treatment follow-up data). Staging was retrospectively assessed and may be subject to misclassification. However, patients with metastatic disease at diagnosis were specifically documented and excluded from PSM to ensure accuracy of comparative analyses.
Clinical data were extracted from institutional electronic health records and validated manually following Eastern Cooperative Oncology Group (ESMO) Guidance for Reporting Oncology real-World evidence (GROW) principles (Supplementary Material, available at https://doi.org/10.1016/j.esmorw.2025.100161). Missing data were handled through listwise deletion for variables used in PSM.
Demographic, clinical, and pathological data were summarized with descriptive statistics [frequencies for categorical variables, median and interquartile ranges (IQR) for continuous ones]. Median follow-up was estimated with reverse Kaplan–Meier method. Survival data were measured from the date of diagnosis, and the events were defined as follows: overall survival (OS), death; disease-free survival (DFS), death or disease recurrence at any site (local and/or regional and/or distant); distant metastasis-free survival (DMFS), death or distant recurrence independent of locoregional status; distant metastasis-free interval (DMFI), distant recurrence independent of locoregional status and death (i.e. dead patients without evidence of distant disease were censored at the date of death); locoregional disease-free survival (LRFS), death or locoregional recurrence independent of distant status; and locoregional control (LRC), locoregional recurrence independent of survival status and metastatic disease (e.g. patients who died with no evidence of locoregional disease were censored). All time-to-event variables were estimated with Kaplan–Meier method, and hazard ratios (HRs) were estimated with Cox proportional hazards model.
The exposure was concomitant CT (i.e. control cases were subjects receiving RT without concurrent CT).
Continuous variables were compared using Mann–Whitney test, and contingency tables with Fisher’s exact test or chi-square test, as appropriate. Survivals were compared with log-rank test.
After analyzing the eligible study population, we focused on patients without metastatic disease at diagnosis and conducted a PSM analysis to minimize the impact of confounding variables on the estimated treatment effects. The propensity scores were estimated using a logistic regression model with the treatment variable as the dependent variable (concomitant CT yes versus no), and considering as covariates those variables with possible prognostic impact and without missing values: sex (women versus men),7 age (continuous variable),9, 10, 11 comorbidities [Adult Comorbidity Evaluation (ACE)-27],12 primary site (major versus minor SGC),9,13 and surgery (yes versus no).1,2,4
To assess the coherence of our findings, we carried out a comprehensive sensitivity analysis using several approaches, which are detailed in the Supplementary Material, available at https://doi.org/10.1016/j.esmorw.2025.100161.
PSM was carried out using the ‘MatchIt’ package, balance checks were conducted using the ‘cobalt’ package, and Rosenbaum bounds sensitivity analysis was carried out using the ‘rbounds’ package. Inverse probability weighting was implemented with the ‘survey’ package. All PSM and sensitivity analyses were conducted using R (R version 4.2.3, 15 March 2023).14
The resulting matched dataset was then analyzed to assess differences in the distribution of clinical variables [sex; age; comorbidities; PS; primary site; pathological type; stage; surgery; resection; regional lymph node involvement with extracapsular spread (ECS); perineural invasion (PNI); lymphovascular infiltration (LVI); RT setting, energy, and dose] and in oncologic outcomes (follow-up, OS, DFS, DMFS, LRFS, LRC). In the matched population, prespecified endpoints were: (i) the absence of significant differences in clinical characteristics and follow-up of patients receiving CRT versus exclusive RT; (ii) to assess any differences in terms of oncologic outcomes in patients receiving RT versus CRT. To quantify the potential benefit of CRT in terms of survival improvement, we analyzed the restricted mean survival time difference at a landmark using the smallest value among the largest observed times across the strata as the tau value in the formulas. No post hoc analyses were conducted. All statistical analyses were carried out using SAS® OnDemand for Academics. Statistical significance was set at 0.05.
All patients were treated following a multidisciplinary discussion including head and neck surgeons, radiation oncologists, medical oncologists, radiologists, and pathologists. After treatment conclusion, patients were followed up as per clinical practice and international guidelines.4 The clinical research question was discussed at multidisciplinary level, and the study design was conceived in an interdisciplinary setting including information technology professionals expert in the management of DWH, electronic case report forms (eCRFs), and complex databases.
Results
Patient characteristics
After case extraction from the institutional DWH, patient selection based on eligibility criteria, and data cleaning, a total of 178 subjects were included in the analysis (Figure 1).
Figure 1.
Flowchart reporting the case selection for analysis. ACC, adenoid cystic carcinoma; CIRT, carbon ion radiotherapy; DWH, data warehouse; H&N, head and neck; RT, radiotherapy.
Most patients (n = 102, 57%) were women, median age was 52 years [interquartile range (IQR) 20 years], ACE-27 was 0-1 in 95%, and Eastern Cooperative Oncology Group (ECOG) performance status (PS) was 0-1 in 98% of cases. Surgery was carried out in 160 patients (89%). RT was delivered with intensity-modulated proton therapy (IMPT) at the National Center for Oncological Hadrontherapy (CNAO) in 23 case; all the remaining subjects (88%) received photon-based IMRT techniques. CRT was administered in 24 patients (13%). Twenty-two (12%) subjects had metastatic disease at diagnosis (20 lung, 2 bone), and after multidisciplinary board discussion received curative RT (CRT in 4 cases).
Median age was 47.5 years (IQR 12.5 years) in patients receiving CRT versus 53 years (IQR 21 years) in those receiving exclusive RT (P = 0.016). In the same groups, median RT dose was 66 Gy (GyE for proton therapy) (IQR 3.95 Gy) versus 64.4 Gy (IQR 6 Gy), respectively (P = 0.0016).
Further clinical characteristics of the cases included in the study and the clinicodemographics and tumor characteristics of patients receiving either exclusive RT or concomitant CRT are detailed in Table 1.
Table 1.
Clinical characteristics of the study population (unmatched)
| Overall, n (%) (n = 178) |
No CT, n (%) (n = 154) |
Concomitant CT, n (%) (n = 24) |
P value | |
|---|---|---|---|---|
| Sex | 0.269 | |||
| F | 102 (57) | 91 (59) | 11 (46) | |
| M | 76 (43) | 63 (41) | 13 (54) | |
| Age, median years (IQR) | 52 (20) | 53 (21) | 47.5 (12.5) | 0.016 |
| ACE-27 | 1.0 (0-1 versus 2-3) | |||
| 0 | 105 (72) | 87 (71) | 18 (75) | |
| 1 | 34 (23) | 29 (24) | 5 (21) | |
| 2 | 6 (4) | 5 (4) | 1 (4) | |
| 3 | 2 (1) | 2 (1) | — | |
| Unknown | 31 | 31 | — | |
| ECOG PS | 0.461 (0 versus 1-2) | |||
| 0 | 80 (81) | 70 | 10 | |
| 1 | 17 (17) | 13 | 4 | |
| 2 | 2 (2) | 2 | — | |
| Unknown | 79 | 69 | 10 | |
| Primary site | 0.829 | |||
| Major SGC | 86 (48) | 75 (49) | 11 (46) | |
| Minor SGC | 92 (52) | 79 (51) | 13 (54) | |
| Pathological type | 0.758 | |||
| Solid | 47 (50) | 42 (51) | 5 (42) | |
| Nonsolid | 47 (50) | 40 (49) | 7 (58) | |
| Unknown | 84 | 72 | 12 | |
| Stage (AJCC/UICC 8th edition) | 0.751 (I-II versus III-IV) | |||
| I | 11 (11) | 9 (11) | 2 (12) | |
| II | 10 (10) | 8 (9) | 2 (12) | |
| III | 22 (22) | 19 (23) | 3 (17) | |
| IV | 57 (57) | 47 (57) | 10 (59) | |
| Unknown | 78 | 71 | 7 | |
| cM1 | 22 (12) | 18 (12) | 4 (17) | |
| Lung | 20 | 17 | 3 | |
| Bone | 2 | 1 | 1 | |
| Surgery | 0.714 | |||
| Carried out | 160 (89) | 139 (90) | 21 (88) | |
| Not carried out | 18 (11) | 15 (10) | 3 (12) | |
| Resection | 0.35 | |||
| R0 | 25 (23) | 23 (25) | 2 (12) | |
| R1/R2 | 82 (77) | 68 (75) | 14 (88) | |
| Unknown | 53 | 48 | 5 | |
| ECS | 0.7 | |||
| Present | 12 (19) | 9 (18) | 3 (23) | |
| Absent | 50 (81) | 40 (82) | 10 (77) | |
| Unknown | 98 | 90 | 8 | |
| PNI | 0.349 | |||
| Present | 96 (91) | 79 (89) | 17 (100) | |
| Absent | 9 (9) | 9 (10) | — | |
| Unknown | 55 | 51 | 4 | |
| LVI | 0.747 | |||
| Present | 25 (41) | 20 (40) | 5 (45) | |
| Absent | 36 (59) | 30 (60) | 6 (55) | |
| Unknown | 99 | 89 | 10 | |
| RT setting | 0.33 | |||
| Adjuvant | 154 (87) | 135 (88) | 19 (79) | |
| Definitive | 24 (13) | 19 (12) | 5 (21) | |
| RT type | 0.317 | |||
| Photons | 156 (88) | 133 (86) | 23 (96) | |
| Protons | 23 (12) | 22 (14) | 1 (4) | |
| RT dose, median Gy (IQR) | 66 (6) | 64.4 (6) | 66 (3.95) | 0.0016 |
| Concomitant CT | 154 | 24 | — | |
| Yes | 24 (13) | |||
| No | 154 (87) |
ACE-27, Adult Comorbidity Evaluation; CT, chemotherapy; ECOG PS, Eastern Cooperative Oncology Group performance status; ECS, extracapsular spread; IQR, interquartile range; LVI, lymphovascular infiltration; PNI, perineural invasion; RT, radiotherapy; SGC, salivary gland carcinoma.
Oncologic outcomes
Median follow-up was 85.2 months (95% CI 78.9-93.9 months): 84 months (95% CI 75-93.9 months) in subjects receiving RT versus 100 months (95% CI 77.7-122.5 months) in those treated with CRT (P = 0.67).
No statistically significant differences were observed in OS, DFS, LRFS, and LRC (Table 2). Median DMFS (Figure 1A) was 101.38 months [95% CI 31.5 months-not reached (NR)] in the 20 patients receiving CRT versus 50.16 months (95% CI 39.04-63.02 months) in the remaining 136 treated with exclusive RT (P = 0.052). Median DMFIs in the same groups were 101.38 months (95% CI 44.07 months-NR) and 53.25 months (95% CI 39.8-65 months), respectively (P = 0.071).
Table 2.
Oncologic outcomes of the study population (unmatched)
| Overall, months (95% CI) (n = 178) |
No CT, months (95% CI) (n = 154) |
Concomitant CT, months (95% CI) (n = 24) |
P value | |
|---|---|---|---|---|
| Median OS | 145.4 (127.8-163.4) | 141.5 (118.1-163.4) | NR (101.3-NR) | 0.627 |
| Median DFS | 37.4 (30.8-42.5) | 37.4 (30.4-40.9) | 42 (19.3-59.4) | 0.291 |
| Median DMFS | 52.86 (40.06-63.25) [n = 156] | 50.16 (39.04-63.02) [n = 136] | 101.38 (31.5-NR) [n = 20] | 0.052 |
| Median LRFS | 81.1 (63-108) | 82 (64.6-117.2) | 56.8 (25.8-NR) | 0.556 |
| Median LRC | NR (70.9-NR) | NR (71-NR) | 67.3 (25.8-NR) | 0.488 |
| Median DMFI | 56.87 (41.05-70.03) | 53.25 (39.8-65) | 101.38 (44.07-NR) | 0.071 |
CI, confidence interval; CT, chemotherapy; DFS, disease-free survival; DMFI, distant metastasis-free interval; DMFS, distant metastasis-free survival; LRC, locoregional control; LRFS, locoregional disease-free survival; NR, not reached; OS, overall survival.
In the unmatched population, 5-year and 10-year-DMFS were 44.1% and 22.8%, respectively. At the same landmarks, the respective DFMIs were 46.8% and 26.5% (Figure 2A). With CRT, the HR for distant failure or death was 0.57 (95% CI 0.32-1.01, P = 0.057) and HR for distant relapse was 0.55 (95% CI 0.28-1.06, P = 0.076). Restricted mean DMFI difference at 168.5 months was 28.1 months (P = 0.068): 99.44 months (CRT) versus 71.28 months (RT).
Figure 2.
Summary of covariate balance of the study populations before (red) and after (green) propensity score matching. M, male; SGC, salivary gland carcinoma.
Toxicity and salvage treatment data were not uniformly recorded across the study period and could not be included systematically. However, no CTCAE v5.0 grade ≥4 toxicities or treatment-related deaths were reported.
Propensity score matching
After PSM, 40 patients were eligible for the planned analysis. Summary of balance for matched data (Figure 3), sample sizes, balance measures and quantile–quantile (QQ) plots, clinical characteristics, follow-up, and oncologic outcomes of the matched patient cohorts are detailed in the Supplementary Material (Tables S1-S6 and Figures S1-S3, available at https://doi.org/10.1016/j.esmorw.2025.100161). Notably, after PSM adjustments, no statistically significant differences were observed in terms of age (median age 43 years versus 47.5 years, P = 0.444), comorbidities (P = 0.694), PS (P = 0.306), and other key variables (Supplementary Table S4, available at https://doi.org/10.1016/j.esmorw.2025.100161).
Figure 3.
Distant metastasis-free interval (DMFI) of the unmatched (A) and matched (B) study cohorts stratified according to chemotherapy (CT) administration (red, with CT; blue, without CT).
OS, DFS, LRFS, and LRC did not differ significantly between the two groups (Supplementary Table S6, available at https://doi.org/10.1016/j.esmorw.2025.100161). Median DMFS (Figure 1B) was 101.3 months (95% CI 31.5 months-NR) versus 39.8 months (95% CI 34-81.1 months) in subjects receiving or not receiving CRT, respectively (P = 0.059). In patients treated with CRT, HR for distant failure or death was 0.48 (95% CI 0.22-1.04, P = 0.064). Median DMFI was 101.38 months (95% CI 44.07 months-NR) in subjects receiving CRT versus 39.8 months (95% CI 34.07-81.1 months) in patients receiving RT (P = 0.032). HR for occurrence of distant metastasis was 0.43 (95% CI 0.19-0.95, P = 0.037) for patients receiving CRT.
In the matched cohorts, restricted mean DMFI difference at 137 months was 28.2 months (89.58 months with CT versus 61.29 months without) for patients receiving CT (Figure 2B, P = 0.059).
Sensitivity analyses confirmed that results remained statistically significant (Supplementary Material Section 4.4, available at https://doi.org/10.1016/j.esmorw.2025.100161), suggesting a moderate level of hidden bias would be required to nullify the observed effect.
Discussion
To our knowledge, this analysis represents the largest study to date comparing survival outcomes between CRT and RT alone in patients with ACC treated with curative intent. The strengths of the current study are that the population is restricted to ACC and there was a long median follow-up. In this study, we also included ACC of the small salivary glands, contributing our real-world insights on this even rarer entity. Additionally, all the patients involved in this analysis received IMRT, and we also included data on proton therapy.
To date, this is the first study to explore the impact of CRT on metastatic progression for ACC patients treated with curative intent. Our results suggest that patients treated with CRT have a lower risk of developing distant metastases and experience slower progression to metastatic disease over time, leading to prolonged metastasis-free survival, even though this does not translate into a longer OS. Specifically, in the present study, we observed a trend indicating a better DMFS (P = 0.052) and a longer DMFI (P = 0.071) in ACC patients treated with CRT. After PSM, the resulting cohorts were well balanced in terms of baseline characteristics (sex, age, comorbidities, primary site, surgery, pathological features including ECS and surgical margins). After PSM, the improvements observed in DMFS and DMFI not only persisted but also became more pronounced, achieving statistical significance (P = 0.017 and P = 0.032, respectively). With CRT, the HR for distant failure or death was 0.57 (P = 0.057) and the HR for distant relapse was 0.55 (P = 0.076).15 These findings were confirmed in the matched cohort, where the HR for the occurrence of distant metastasis reached statistical significance (P = 0.037). Furthermore, in our matched population, the restricted mean DMFI difference at 137 months was 28.2 months for patients receiving CT (P = 0.059). Given the indolent nature of ACC and the predominant risk of late distant metastasis, DMFI was selected as the primary endpoint to assess the potential delay in metastatic spread. OS differences may not emerge within the timeframe of available follow-up.
Regarding oncologic outcomes, in our study, no statistically significant differences were observed in OS, DFS, LRFS, and LRC. To date, no other retrospective study demonstrated a significant survival benefit in OS, DSF, and PFS in patients treated with concurrent CRT versus RT alone. In the two largest retrospective studies, Tanvetyanon et al. (741 patients)16 and Amini et al. (2210 patients)17 found no significant OS benefit with CRT in the adjuvant setting. Notably, in the study by Amini et al., this finding persisted even under PSM.17 Furthermore, both studies reported that concurrent treatment was associated with increased mortality.16,17 It should be noted that the study by Tanvetyanon et al. included patients aged 66 years or older,16 and both studies also included histological subtypes with more aggressive behavior than ACC.16,17
SGCs are considered a rare malignancy, and conducting randomized clinical trials is difficult. The evidence supporting the combination of CRT for SGC, including ACC, is derived from retrospective trials and remains limited.4,5 Therefore, combination therapy is not recommended for patients receiving adjuvant RT and those undergoing definitive RT for nonoperable tumors, except in clinical trials. Previous studies have investigated the effect of CRT in patients with locally advanced SGC. However, only a few studies have specifically focused on ACC, which can differ from other histologies in both its biological and clinical behavior and consequently in prognosis. Moreover, only one study has investigated the use of CRT in both patients receiving either post-operative RT or RT for unresectable disease.18
Similarly, Mifsud et al. evaluated DMFS but in a cohort of patients with diverse tumor histologies: in their study, the poorer DMFS observed in the CRT group compared with patients treated with RT alone may have been influenced by the higher recurrence rate for salivary duct carcinomas (65%) compared with that for ACC (28%).19 In a retrospective study of patients with operated salivary gland ACC, Hsieh et al. demonstrated a trend toward higher locoregional control by delivering CRT after surgery.20 This trend remained statistically significant after PSM and was particularly pronounced in patients with positive surgical margins, perineural invasion, and stage III-IV disease (AJCC/UICC 8th edition). Indeed, no significant differences between CRT and RT were observed in terms of DFS and OS, according to our findings. In the same study, the difference between the groups in DMFS was not statistically significant in the entire cohort (P = 0.123) and was even weaker in the matched population (P = 0.656). Furthermore, in their analysis, distant metastases represented the most frequent treatment failure and the most common cause of disease-specific mortality. In our population, the estimated 10-year DMFI was 26.5%, highlighting the relevant risk of distant failure even at prolonged follow-up.
Rosenberg et al. and Schoenfeld et al. demonstrated that CRT was well tolerated with expected toxicities, and no treatment-related deaths occurred in SGC patients.18,21 Schoenfeld et al. also found excellent local control rates in patients with high-risk features, with no patient developing regional failure at a median follow-up of 2.3 years and only one patient experiencing local failure.21 In the definitive setting, three papers consistently reported better local control outcomes with CRT compared with definitive RT.22, 23, 24
Surgery was included as a key variable for PSM. Indeed, RT setting (adjuvant versus definitive) was evaluated in the matched cohort and did not significantly differ between groups. Due to limited numbers, subgroup analyses by RT setting were not carried out but should be addressed in future studies.
The median follow-up period in our study was 85.2 months (95% CI 78.9-93.9 months), which stands out as the longest reported in the available studies to date. Considering the indolent course of ACC, a long-term follow-up is essential for obtaining reliable data; therefore, it is known that distant metastases can occur even 10-20 years after diagnosis.25 The observational period of our study was limited to 2007-2022 to ensure the inclusion of patients treated with IMRT. Additionally, our cohort included 23 patients who received IMPT.
Due to the retrospective analysis of the study, the present work has several limitations. Firstly, some key variables (e.g. perineural invasion, lymphovascular infiltration) had incomplete records in a non-negligible fraction of patients (e.g. 28% missing resection margin status). Since variables with missing data were not included as potential in the PSM, we cannot exclude that, if these variables were fully available, the final analysis could have led to a different matched population, possibly with different results.
Another limitation is possible selection bias. Given that patients lost to follow-up were excluded, survival estimates may have been skewed. However, the reason for excluding them was essentially related to the scarcity of data in some cases, which could have led to additional missing data possibly negatively interfering with the models.
Genomic and socioeconomic data were unavailable for most patients. While this may introduce unmeasured confounding, sensitivity analyses support the robustness of our conclusions.
The single-center setting may limit external generalizability, although the institutional expertise and standardized multidisciplinary management reflect a real-world referral scenario that may inform clinical decisions in comparable high-volume centers.
The limited number of patients treated with CRT may impact statistical power. However, multiple sensitivity analyses and covariate balancing confirmed the consistency of our findings, supporting their internal validity despite the relatively small matched sample, which is nonetheless meaningful for such a rare disease.
Conclusions
Despite these limitations and the small size of our cohort, a reproducible statistical model showed a significant improvement in DMFS and DMFI in patients treated with CRT. These findings suggest that CRT may significantly delay the onset of distant metastases in ACC patients managed with curative intent. Thus, the opportunity to intensify the locoregional treatments of ACC patients by delivering CRT should be carefully taken into consideration, especially in young and fit subjects.
Acknowledgements
The conducting of the research activities reported in the article, the collection, analysis and interpretation of data, and the preparation of the article and its submission were not funded.
Funding
This research was partially funded by Italian Ministry of Health “Ricerca Corrente” funds and by AIRC (IG29083 to L.L).
Disclosure
The first author (SC) discloses occasional fees for participation as a speaker at conferences/congresses from AccMed and support for attending meetings and/or travel from AccMed, MultiMed Engineers srl, and Care Insight sas, unrelated to the content of this work.
The last author (LL) discloses the following conflicts of interest: research funds donated directly to the institution for clinical trials from AstraZeneca, Bristol Myers Squibb, Boehringer Ingelheim, Celgene International, Eisai, Exelixis, Debiopharm International SA, Hoffmann-La Roche Ltd, IRX Therapeutics, Medpace, Merck-Serono, Merck Sharpe & Dohme (MSD), Novartis, Pfizer, Roche, and Buran and occasional fees for participation as a speaker at conferences/congresses or as a scientific consultant for advisory boards from AstraZeneca, Bayer, MSD, Merck-Serono, AccMed, Neutron Therapeutics, Inc., and Alentis.
The remaining authors have declared no conflicts of interest.
All the authors declare that the research reported in the current article was conducted without any commercial or financial relationships that could be construed as a potential conflict of interest.
Supplementary data
References
- 1.Bishop J.A., Thompson L.D.R., Wakely P.E., Jr., Weinreb I. Vol. 5. ARP Press; Rockville, MD: 2021. AFIP Atlas: Tumors of the Salivary Glands. (Fifth Series). [Google Scholar]
- 2.WHO Classification of Tumours Editorial Board . 5th ed. Vol. 9. World Health Organization; Geneva, Switzerland: 2024. Head and Neck Tumours—WHO Classification of Tumours. [Google Scholar]
- 3.RARECARENet—Data Source and Methods. http://www.rarecarenet.eu/ Available at.
- 4.Geiger J.L., Ismaila N., Beadle B., et al. Management of salivary gland malignancy: ASCO guideline. J Clin Oncol. 2021;39(17):1909–1941. doi: 10.1200/JCO.21.00449. [DOI] [PubMed] [Google Scholar]
- 5.van Herpen C., Vander Poorten V., Skalova A., et al. Salivary gland cancer: ESMO–European Reference Network on Rare Adult Solid Cancers (EURACAN) Clinical Practice Guideline for diagnosis, treatment and follow-up. ESMO Open. 2022;7(6) doi: 10.1016/j.esmoop.2022.100602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Trama A., Licitra L., Cavalieri S., et al. The observational clinical registry (cohort design) of the European Reference Network on Rare Adult Solid Cancers: the protocol for the rare head and neck cancers. PLoS One. 2023;18(3) doi: 10.1371/journal.pone.0283071. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Cavalieri S., Mariani L., Vander Poorten V., et al. Prognostic nomogram in patients with metastatic adenoid cystic carcinoma of the salivary glands. Eur J Cancer. 2020;136:35–42. doi: 10.1016/j.ejca.2020.05.013. [DOI] [PubMed] [Google Scholar]
- 8.Hammami L., Paglialonga A., Pruneri G., et al. Automated classification of cancer morphology from Italian pathology reports using Natural Language Processing techniques: a rule-based approach. J Biomed Inform. 2021;116 doi: 10.1016/j.jbi.2021.103712. [DOI] [PubMed] [Google Scholar]
- 9.Amit M., Binenbaum Y., Sharma K., et al. Analysis of failure in patients with adenoid cystic carcinoma of the head and neck. An international collaborative study. Head Neck. 2014;36(7):998–1004. doi: 10.1002/hed.23405. [DOI] [PubMed] [Google Scholar]
- 10.Jang S., Patel P.N., Kimple R.J., McCulloch T.M. Clinical outcomes and prognostic factors of adenoid cystic carcinoma of the head and neck. Anticancer Res. 2017;37(6):3045. doi: 10.21873/anticanres.11659. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lloyd S., Yu J.B., Wilson L.D., Decker R.H. Determinants and patterns of survival in adenoid cystic carcinoma of the head and neck, including an analysis of adjuvant radiation therapy. Am J Clin Oncol. 2011;34(1):76–81. doi: 10.1097/COC.0b013e3181d26d45. [DOI] [PubMed] [Google Scholar]
- 12.Piccirillo J.F., Tierney R.M., Costas I., Grove L., Spitznagel E.L. Prognostic importance of comorbidity in a hospital-based cancer registry. JAMA. 2004;291(20):2441–2447. doi: 10.1001/jama.291.20.2441. [DOI] [PubMed] [Google Scholar]
- 13.Tasoulas J., Schrank T.P., Bharambe H., et al. Molecular characterization of the salivary adenoid cystic carcinoma immune landscape by anatomic subsites. Sci Rep. 2024;14(1):1–13. doi: 10.1038/s41598-024-66709-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.R: The R Project for Statistical Computing. https://www.r-project.org/ Available at.
- 15.Kang N.W., Kuo Y.H., Wu H.C., Ho C.H., Chen Y.C., Yang C.C. No survival benefit from adding chemotherapy to adjuvant radiation in advanced major salivary gland cancer. Sci Rep. 2022;12(1):1–10. doi: 10.1038/s41598-022-25468-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Tanvetyanon T., Fisher K., Caudell J., Otto K., Padhya T., Trotti A. Adjuvant chemoradiotherapy versus with radiotherapy alone for locally advanced salivary gland carcinoma among older patients. Head Neck. 2016;38(6):863–870. doi: 10.1002/hed.24172. [DOI] [PubMed] [Google Scholar]
- 17.Amini A., Waxweiler T.V., Brower J.V., et al. Association of adjuvant chemoradiotherapy vs radiotherapy alone with survival in patients with resected major salivary gland carcinoma: data from the National Cancer Data Base. JAMA Otolaryngo Head Neck Surg. 2016;142(11):1100–1110. doi: 10.1001/jamaoto.2016.2168. [DOI] [PubMed] [Google Scholar]
- 18.Rosenberg L., Weissler M., Hayes D.N., et al. Concurrent chemoradiotherapy for locoregionally advanced salivary gland malignancies. Head Neck. 2012;34(6):872–876. doi: 10.1002/hed.21831. [DOI] [PubMed] [Google Scholar]
- 19.Mifsud M.J., Tanvetyanon T., Mccaffrey J.C., et al. Adjuvant radiotherapy versus concurrent chemoradiotherapy for the management of high-risk salivary gland carcinomas. Head Neck. 2016;38(11):1628–1633. doi: 10.1002/hed.24484. [DOI] [PubMed] [Google Scholar]
- 20.Hsieh C.E., Lin C.Y., Lee L.Y., et al. 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. Radiat Oncol. 2016;11:47. doi: 10.1186/s13014-016-0617-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Schoenfeld J.D., Sher D.J., Norris C.M., et al. Salivary gland tumors treated with adjuvant intensity-modulated radiotherapy with or without concurrent chemotherapy. Int J Radiat Oncol Biol Phys. 2012;82(1):308–314. doi: 10.1016/j.ijrobp.2010.09.042. [DOI] [PubMed] [Google Scholar]
- 22.Bhattasali O., Holliday E., Kies M.S., et al. Definitive proton radiation therapy and concurrent cisplatin for unresectable head and neck adenoid cystic carcinoma: a series of 9 cases and a critical review of the literature. Head Neck. 2016;38(suppl 1):E1472–E1480. doi: 10.1002/hed.24262. [DOI] [PubMed] [Google Scholar]
- 23.Samant S., Van Den Brekel M.W., Kies M.S., et al. Concurrent chemoradiation for adenoid cystic carcinoma of the head and neck. Head Neck. 2012;34(9):1263–1268. doi: 10.1002/hed.21905. [DOI] [PubMed] [Google Scholar]
- 24.Haddad R.I., Posner M.R., Busse P.M., et al. Chemoradiotherapy for adenoid cystic carcinoma: preliminary results of an organ sparing approach. Am J Clin Oncol. 2006;29(2):153–157. doi: 10.1097/01.coc.0000203756.36866.17. [DOI] [PubMed] [Google Scholar]
- 25.Terhaard C.H.J., Lubsen H., Van Der Tweel I., et al. Salivary gland carcinoma: independent prognostic factors for locoregional control, distant metastases, and overall survival: results of the Dutch head and neck oncology cooperative group. Head Neck. 2004;26(8):681–693. doi: 10.1002/hed.10400. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.