Artificial intelligence-assisted quantification of COVID-19 pneumonia burden from computed tomography improves prediction of adverse outcomes over visual scoring systems.

OBJECTIVE: We aimed to evaluate the effectiveness of utilizing artificial intelligence (AI) to quantify the extent of pneumonia from chest CT scans, and to determine its ability to predict clinical deterioration or mortality in patients admitted to the hospital with COVID-19 in comparison to semi-quantitative visual scoring systems. METHODS: A deep-learning algorithm was utilized to quantify the pneumonia burden, while semi-quantitative pneumonia severity scores were estimated through visual means. The primary outcome was clinical deterioration, the composite end point including admission to the intensive care unit, need for invasive mechanical ventilation, or vasopressor therapy, as well as in-hospital death. RESULTS: The final population comprised 743 patients (mean age 65  ±â€¯ 17 years, 55% men), of whom 175 (23.5%) experienced clinical deterioration or death. The area under the receiver operating characteristic curve (AUC) for predicting the primary outcome was significantly higher for AI-assisted quantitative pneumonia burden (0.739, p = 0.021) compared with the visual lobar severity score (0.711, p < 0.001) and visual segmental severity score (0.722, p = 0.042). AI-assisted pneumonia assessment exhibited lower performance when applied for calculation of the lobar severity score (AUC of 0.723, p = 0.021). Time taken for AI-assisted quantification of pneumonia burden was lower (38 ± 10 s) compared to that of visual lobar (328 ± 54 s, p < 0.001) and segmental (698 ± 147 s, p < 0.001) severity scores. CONCLUSION: Utilizing AI-assisted quantification of pneumonia burden from chest CT scans offers a more accurate prediction of clinical deterioration in patients with COVID-19 compared to semi-quantitative severity scores, while requiring only a fraction of the analysis time. ADVANCES IN KNOWLEDGE: Quantitative pneumonia burden assessed using AI demonstrated higher performance for predicting clinical deterioration compared to current semi-quantitative scoring systems. Such an AI system has the potential to be applied for image-based triage of COVID-19 patients in clinical practice.

Imaging of Cystic Lung Disease.

Diffuse cystic lung disease refers to multiple rounded lucencies or low-attenuating areas with well-defined interfaces with normal lung. Parenchymal lucencies, such as cavitary disease, may mimic cystic lung disease. Cystic lung disease generally has a nonspecific presentation. Pulmonary cysts may present in isolation or with ancillary imaging features, such as ground-glass opacities or nodules. Clinical features, such as connective tissue disease, can narrow the differential diagnosis. In cases with indeterminate imaging and clinical features, open lung biopsy should be considered. Ultrarare cystic lung diseases, such as light chain deposition disease, can mimic more common diseases.

Rapid quantification of COVID-19 pneumonia burden from computed tomography with convolutional long short-term memory networks.

Purpose: Quantitative lung measures derived from computed tomography (CT) have been demonstrated to improve prognostication in coronavirus disease 2019 (COVID-19) patients but are not part of clinical routine because the required manual segmentation of lung lesions is prohibitively time consuming. We aim to automatically segment ground-glass opacities and high opacities (comprising consolidation and pleural effusion). Approach: We propose a new fully automated deep-learning framework for fast multi-class segmentation of lung lesions in COVID-19 pneumonia from both contrast and non-contrast CT images using convolutional long short-term memory (ConvLSTM) networks. Utilizing the expert annotations, model training was performed using five-fold cross-validation to segment COVID-19 lesions. The performance of the method was evaluated on CT datasets from 197 patients with a positive reverse transcription polymerase chain reaction test result for SARS-CoV-2, 68 unseen test cases, and 695 independent controls. Results: Strong agreement between expert manual and automatic segmentation was obtained for lung lesions with a Dice score of 0.89 ± 0.07 ; excellent correlations of 0.93 and 0.98 for ground-glass opacity (GGO) and high opacity volumes, respectively, were obtained. In the external testing set of 68 patients, we observed a Dice score of 0.89 ± 0.06 as well as excellent correlations of 0.99 and 0.98 for GGO and high opacity volumes, respectively. Computations for a CT scan comprising 120 slices were performed under 3 s on a computer equipped with an NVIDIA TITAN RTX GPU. Diagnostically, the automated quantification of the lung burden % discriminate COVID-19 patients from controls with an area under the receiver operating curve of 0.96 (0.95-0.98). Conclusions: Our method allows for the rapid fully automated quantitative measurement of the pneumonia burden from CT, which can be used to rapidly assess the severity of COVID-19 pneumonia on chest CT.

Rapid quantification of COVID-19 pneumonia burden from computed tomography with convolutional LSTM networks.

Quantitative lung measures derived from computed tomography (CT) have been demonstrated to improve prognostication in Coronavirus disease 2019 (COVID-19) patients, but are not part of the clinical routine since required manual segmentation of lung lesions is prohibitively time-consuming. We propose a new fully automated deep learning framework for quantification and differentiation between lung lesions in COVID-19 pneumonia from both contrast and non-contrast CT images using convolutional Long Short-Term Memory (LSTM) networks. Utilizing the expert annotations, model training was performed using 5-fold cross-validation to segment ground-glass opacity and high opacity (including consolidation and pleural effusion). The performance of the method was evaluated on CT data sets from 197 patients with positive reverse transcription polymerase chain reaction test result for SARS-CoV-2. Strong agreement between expert manual and automatic segmentation was obtained for lung lesions with a Dice score coefficient of 0.876 ± 0.005; excellent correlations of 0.978 and 0.981 for ground-glass opacity and high opacity volumes. In the external validation set of 67 patients, there was dice score coefficient of 0.767 ± 0.009 as well as excellent correlations of 0.989 and 0.996 for ground-glass opacity and high opacity volumes. Computations for a CT scan comprising 120 slices were performed under 2 seconds on a personal computer equipped with NVIDIA Titan RTX graphics processing unit. Therefore, our deep learning-based method allows rapid fully-automated quantitative measurement of pneumonia burden from CT and may generate the big data with an accuracy similar to the expert readers.

Epicardial adipose tissue is associated with extent of pneumonia and adverse outcomes in patients with COVID-19.

AIM: We sought to examine the association of epicardial adipose tissue (EAT) quantified on chest computed tomography (CT) with the extent of pneumonia and adverse outcomes in patients with coronavirus disease 2019 (COVID-19). METHODS: We performed a post-hoc analysis of a prospective international registry comprising 109 consecutive patients (age 64 ±â€¯16 years; 62% male) with laboratory-confirmed COVID-19 and noncontrast chest CT imaging. Using semi-automated software, we quantified the burden (%) of lung abnormalities associated with COVID-19 pneumonia. EAT volume (mL) and attenuation (Hounsfield units) were measured using deep learning software. The primary outcome was clinical deterioration (intensive care unit admission, invasive mechanical ventilation, or vasopressor therapy) or in-hospital death. RESULTS: In multivariable linear regression analysis adjusted for patient comorbidities, the total burden of COVID-19 pneumonia was associated with EAT volume (ß = 10.6, p = 0.005) and EAT attenuation (ß = 5.2, p = 0.004). EAT volume correlated with serum levels of lactate dehydrogenase (r = 0.361, p = 0.001) and C-reactive protein (r = 0.450, p < 0.001). Clinical deterioration or death occurred in 23 (21.1%) patients at a median of 3 days (IQR 1-13 days) following the chest CT. In multivariable logistic regression analysis, EAT volume (OR 5.1 [95% CI 1.8-14.1] per doubling p = 0.011) and EAT attenuation (OR 3.4 [95% CI 1.5-7.5] per 5 Hounsfield unit increase, p = 0.003) were independent predictors of clinical deterioration or death, as was total pneumonia burden (OR 2.5, 95% CI 1.4-4.6, p = 0.002), chronic lung disease (OR 1.3 [95% CI 1.1-1.7], p = 0.011), and history of heart failure (OR 3.5 [95% 1.1-8.2], p = 0.037). CONCLUSIONS: EAT measures quantified from chest CT are independently associated with extent of pneumonia and adverse outcomes in patients with COVID-19, lending support to their use in clinical risk stratification.

Quantitative Burden of COVID-19 Pneumonia on Chest CT Predicts Adverse Outcomes: A Post-Hoc Analysis of a Prospective International Registry.

PURPOSE: To examine the independent and incremental value of CT-derived quantitative burden and attenuation of COVID-19 pneumonia for the prediction of clinical deterioration or death. METHODS: This was a retrospective analysis of a prospective international registry of consecutive patients with laboratory-confirmed COVID-19 and chest CT imaging, admitted to four centers between January 10 and May 6, 2020. Total burden (expressed as a percentage) and mean attenuation of ground glass opacities (GGO) and consolidation were quantified from CT using semi-automated research software. The primary outcome was clinical deterioration (intensive care unit admission, invasive mechanical ventilation, or vasopressor therapy) or in-hospital death. Logistic regression was performed to assess the predictive value of clinical and CT parameters for the primary outcome. RESULTS: The final population comprised 120 patients (mean age 64 ± 16 years, 78 men), of whom 39 (32.5%) experienced clinical deterioration or death. In multivariable regression of clinical and CT parameters, consolidation burden (odds ratio [OR], 3.4; 95% confidence interval [CI]: 1.7, 6.9 per doubling; P = .001) and increasing GGO attenuation (OR, 3.2; 95% CI: 1.3, 8.3 per standard deviation, P = .02) were independent predictors of deterioration or death; as was C-reactive protein (OR, 2.1; 95% CI: 1.3, 3.4 per doubling; P = .004), history of heart failure (OR 1.3; 95% CI: 1.1, 1.6, P = .01), and chronic lung disease (OR, 1.3; 95% CI: 1.0, 1.6; P = .02). Quantitative CT measures added incremental predictive value beyond a model with only clinical parameters (area under the curve, 0.93 vs 0.82, P = .006). The optimal prognostic cutoffs for burden of COVID-19 pneumonia as determined by Youden's index were consolidation of greater than or equal to 1.8% and GGO of greater than or equal to 13.5%. CONCLUSIONS: Quantitative burden of consolidation or GGO on chest CT independently predict clinical deterioration or death in patients with COVID-19 pneumonia. CT-derived measures have incremental prognostic value over and above clinical parameters, and may be useful for risk stratifying patients with COVID-19.

Feasibility of optimizing intensity-modulated radiation therapy plans based on measured mucosal dose adjacent to dental fillings and toxicity outcomes.

We prospectively investigated the feasibility of IMRT treatment plan optimization based on dosimeter measurements of lateral tongue mucosal dose adjacent to the dental fillings and evaluated dose-toxicity relationship and factors affecting oral mucositis (OM) in head and neck cancer patients. Twenty-nine head and neck cancer patients with metallic dental fillings who were scheduled to undergo fractionated external beam radiation therapy (RT) ± chemotherapy were enrolled. The lateral tongue dose was measured and if the calculated dose for the entire treatment was ≥35 Gy, a re-plan was generated to reduce the lateral tongue mucosal dose. OM was graded weekly according to Common Terminology Criteria for Adverse Events version 4.0 and the patients completed the Oral Mucositis Weekly Questionnaire-Head and Neck Cancer. The result showed that it was not feasible to optimize the IMRT plan based on measured tongue dose in most of the patients who needed re-plan as re-planning compromised the target coverage in 60% of these patients. The duration of grade (Gr) 2 OM was correlated with measured lateral tongue dose (P = 0.050). Concurrent cetuximab was significantly associated with faster onset of Gr2 OM than concurrent cisplatin (P = 0.006) and with longer duration of OM (P = 0.041) compared to concurrent cisplatin or IMRT-alone. The pattern of reported pain over time was significantly different for each treatment type (RT and cetuximab, RT and cisplatin and RT-alone) and depending on the dose level (P = 0.006). In conclusion, optimizing the IMRT plan based on measured lateral tongue dose was not feasible. Measured lateral tongue dose was significantly correlated with longer duration of OM ≥Gr2, and concurrent cetuximab was associated with earlier onset and longer duration of OM ≥Gr2.

Hearing evaluation of patients with head and neck cancer: Comparison of Common Terminology Criteria for Adverse Events, Brock and Chang adverse event criteria in patients receiving cisplatin.

BACKGROUND: The purpose of this study was to compare Common Terminology Criteria for Adverse Events (CTCAE), Brock and Chang hearing loss grading in patients with head and neck cancer receiving cis-diamminedichloroplatinum (CDDP). Endpoints were baseline distribution of hearing loss, interobserver consistency, and sensitivity to hearing loss after CDDP treatment. METHODS: Four hundred sixty single ear audiograms in 110 patients with head and neck cancer were graded. Hearing loss at baseline, interobserver agreement rates, and changes in hearing loss after CDDP were evaluated. RESULTS: The Chang and Brock tools' baseline hearing loss distribution was concentrated at grade 0 (57% and 41%, respectively), whereas 47%, per the CTCAE, had grade 3 baseline hearing loss. Interobserver agreement was highest for the Brock scale (≥90%) followed by the Chang (≥89%) and CTCAE (≥75%) scales. Detection of change after CDDP was highest for Chang (48%) followed by Brock (45%) and the CTCAE (32%). CONCLUSION: The Brock and Chang tools may be superior to the CTCAE in patients with head and neck cancer receiving CDDP using baseline hearing loss distribution, interobserver agreement, and detection of hearing loss grade change as performance indicators.

Concurrent cetuximab versus platinum-based chemoradiation for the definitive treatment of locoregionally advanced head and neck cancer.

BACKGROUND: The purpose of this study was to present our experience utilizing cetuximab and platinum-based concurrent chemoradiotherapy for the definitive treatment of head and neck squamous cell carcinoma (HNSCC). METHODS: Patients (n = 177) who received definitive concurrent chemoradiotherapy for HNSCC were stratified into 3 groups: receiving cetuximab monotherapy (n = 24), cetuximab and chemotherapy combination (n = 33), or platinum-based chemotherapy without cetuximab (n = 120). Primary endpoints were freedom from relapse, event-free survival, and overall survival (OS). RESULTS: Patients receiving cetuximab monotherapy were older with lower Karnofsky performance status (KPS) and higher Charlson comorbidity scores compared with those treated with combination cetuximab and chemotherapy or platinum-based concurrent chemoradiotherapy. Patients treated with platinum-based concurrent chemoradiotherapy exhibited significantly better freedom from relapse, event-free survival, and OS compared with those receiving cetuximab monotherapy or cetuximab and chemotherapy combination therapies (all p < .05). Differences between patients receiving cetuximab monotherapy and platinum-based concurrent chemoradiotherapy held on multivariate Cox regression. CONCLUSION: This study suggests that platinum-based concurrent chemoradiotherapy is superior to cetuximab-based monotherapy for the definitive treatment of HNSCC.

Galectin-1 mediates radiation-related lymphopenia and attenuates NSCLC radiation response.

PURPOSE: Radiotherapy can result in lymphopenia, which has been linked to poorer survival. Here, we test the hypothesis that radiotherapy-induced lymphopenia is mediated by a tumor-secreted factor, Galectin-1 (Gal-1), which possesses T-cell proapoptotic activities. EXPERIMENTAL DESIGN: Matched Gal-1 wild-type (WT) or null mice were implanted with Lewis lung carcinoma (LLC-1) that either expressed Gal-1 or had Gal-1 stably downregulated. Tumors were irradiated locally and circulating Gal-1 and T cells were measured. Tumor growth, lung metastasis, intratumoral T-cell apoptosis, and microvessel density count were quantified. Thiodigalactoside (TDG), a Gal-1 inhibitor, was used to inhibit Gal-1 function in another group of mice to validate the observations noted with Gal-1 downregulation. Lymphocyte counts, survival, and plasma Gal-1 were analyzed in cohorts of radiotherapy-treated lung [non-small cell lung cancer (NSCLC)] and head and neck cancer patients. RESULTS: LLC irradiation increased Gal-1 secretion and decreased circulating T cells in mice, regardless of host Gal-1 expression. Inhibition of tumor Gal-1 with either shRNA or thiodigalactoside ablated radiotherapy-induced lymphopenia. Irradiated shGal-1 tumors showed significantly less intratumoral CD8(+) T-cell apoptosis and microvessel density, which led to marked tumor growth delay and reduced lung metastasis compared with controls. Similar observations were made after thiodigalactoside treatment. Radiotherapy-induced lymphopenia was associated with poorer overall survival in patients with NSCLC treated with hypofractionated radiotherapy. Plasma Gal-1 increased whereas T-cell decreased after radiation in another group of patients. CONCLUSIONS: Radiotherapy-related systemic lymphopenia appeared to be mediated by radiotherapy-induced tumor Gal-1 secretion that could lead to tumor progression through intratumoral immune suppression and enhanced angiogenesis.

Postradiotherapy CA19-9 kinetics correlate with outcomes in patients with pancreatic adenocarcinoma.

OBJECTIVES: We sought to determine if carbohydrate antigen 19-9 (CA19-9 ) nadir (nCA19-9), time to nadir (TTN), and doubling time (DT) after radiotherapy (RT) correlate with outcomes in pancreatic ductal adenocarcinoma. METHODS: We examined the records of 102 patients treated with RT for primary, nonmetastatic pancreatic ductal adenocarcinoma between August 1998 and July 2011. Of these, 33 patients were treated with postoperative chemoradiotherapy (PORT) and 69 patients with definitive chemoradiotherapy (CRT). RESULTS: Among the patients treated with PORT, TTN and DT were associated with both overall survival (OS; P = <0.01 for both) and freedom from progression (FFP; P = <0.01 for both). In patients treated with CRT, nCA19-9 and TTN correlated with both OS (P = <0.01 and P = 0.02, respectively) and FFP (P = 0.01 and <0.01, respectively). On multivariable analysis, in patients treated with PORT, TTN remained independently correlated with OS and FFP (P = 0.01; hazard ratios [HR], 6.43 and P = 0.02; HR, 4.00, respectively), whereas DT remained independently correlated to FFP (P = 0.04; HR, 0.27). In patients treated with CRT, controlling for pretreatment CA19-9, nCA19-9 and TTN independently correlated with OS (P = <0.01; HR, 3.0 and P = 0.03; HR, 2.56, respectively) and FFP (P = 0.04; HR, 2.31 and P = <0.01; HR, 4.0, respectively). CONCLUSIONS: CA19-9 kinetics after RT correlate with disease progression and survival and could serve as a prognostic tool to guide treatment decisions.

Salvage treatment for locally recurrent nasopharyngeal carcinoma (NPC).

BACKGROUND: It is important to determine the outcomes of retreatment in patients with locally recurrent nasopharyngeal carcinoma. METHODS: We reviewed the records of patients treated for local recurrence at Stanford and Shantou Universities. The end points were local relapse-free survival (LRFS) and overall survival after retreatment. RESULTS: Fifty-six patients from Stanford and 98 from Shantou qualified. For the Stanford patients, 33 had surgery alone (S group), 12 had surgery plus radiotherapy±chemotherapy (CMT group), and 22 had radiotherapy±chemotherapy (RT Stanford group). All Shantou patients received radiotherapy±chemotherapy (RT Shantou group). The 5-year LRFS rates were: 57% for S group, 25% for CMT group, 53% for RT Stanford group, and 41% for RT Shantou group (P>0.05) for rT1-2 tumors; they were 29% for S group, 25% for CMT group, 39% for RT Stanford group, and 9% for RT Shantou group for rT3-4 tumors (P>0.05). For RT patients, 5-year overall survival rates were 49% for Stanford and 25% for Shantou patients (P=0.026). CONCLUSIONS: Similar and durable LRFS rates were attained for both S and RT groups when stratified by rT-stage.

Radiologic assessment of retropharyngeal node involvement in oropharyngeal carcinomas stratified by HPV status.

Radiation of retropharyngeal nodes (RPN) results in increased toxicities. This study assessed characteristics associated with RPN involvement in 165 oropharynx cancer patients. Factors associated with involvement were stage N2c-3 disease and stage N2b disease with either advanced T-stage, ⩾3 involved cervical LN, and ⩾1 involved contralateral LN, or lateral/posterior subsites.

A novel aldehyde dehydrogenase-3 activator (Alda-89) protects submandibular gland function from irradiation without accelerating tumor growth.

PURPOSE: To determine the effect of Alda-89 (an ALDH3 activitor) on (i) the function of irradiated (radiotherapy) submandibular gland (SMG) in mice, (ii) its toxicity profile, and (iii) its effect on the growth of head and neck cancer (HNC) in vitro and in vivo. EXPERIMENTAL DESIGN: Adult mice were infused with Alda-89 or vehicle before, during, and after radiotherapy. Saliva secretion was monitored weekly. Hematology, metabolic profile, and postmortem evaluation for toxicity were examined at the time of sacrifice. Alda-89 or vehicle was applied to HNC cell lines in vitro, and severe combined immunodeficient (SCID) mice transplanted with HNC in vivo with or without radiation; HNC growth was monitored. The ALDH3A1 and ALDH3A2 protein expression was evaluated in 89 patients with HNC and correlated to freedom from relapse (FFR) and overall survival (OS). RESULTS: Alda-89 infusion significantly resulted in more whole saliva production and a higher percentage of preserved acini after radiotherapy compared with vehicle control. There was no difference in the complete blood count, metabolic profile, and major organ morphology between the Alda-89 and vehicle groups. Compared with vehicle control, Alda-89 treatment neither accelerated HNC cell proliferation in vitro, nor did it affect tumor growth in vivo with or without radiotherapy. Higher expression of ALDH3A1 or ALDH3A2 was not significantly associated with worse FFR or OS in either human papillomavirus (HPV)-positive or HPV-negative group. CONCLUSION: Alda-89 preserves salivary function after radiotherapy without affecting HNC growth or causing measurable toxicity in mice. It is a promising candidate to mitigate radiotherapy-related xerostomia.

Long-term outcomes of surgery followed by radiation therapy for minor salivary gland carcinomas.

OBJECTIVES/HYPOTHESIS: Postoperative radiation therapy is often used in patients with high-risk salivary gland carcinomas. In this study we evaluated the outcomes and prognostic factors in patients with minor salivary gland cancers treated with adjuvant radiation therapy. STUDY DESIGN: Retrospective cohort study. METHODS: We performed a retrospective analysis of 90 patients treated with curative intent. Median follow-up was 71 months. Fifty-eight patients (64%) had adenoid cystic carcinomas, 22 (24%) had adenocarcinomas, and 10 (11%) had mucoepidermoid cancers. Primary disease site included 39 (43%) sinonasal, 35 (39%) oral cavity, 10 (11%) oropharynx, and six (7%) others. Twenty-seven patients (30%) were treated with intensity-modulated radiation therapy. RESULTS: Eight local, four neck, and 24 distant relapses were detected. Local control rates at 5 and 10 years were 90% and 88%, respectively. Advanced T stage was associated with worse local control. Distant metastasis rates were 24% and 28% at 5 and 10 years, respectively. Tumor stage, histology, perineural invasion, and lymphovascular space invasion were significant predictors of distant metastasis on univariate analysis. However, on multivariate analysis only the American Joint Committee on Cancer stage was significant. Overall survival rates were 76% and 63% at 5 and 10 years, respectively. More advanced T stage and N stage correlated with worse overall survival. CONCLUSIONS: Tumor stage remains the best predictor for locoregional and distant disease control of minor salivary gland cancers. Postoperative radiation therapy for high-risk patients results in excellent long-term locoregional disease control. Further work is needed to improve systemic control.

Radiotherapy for nonadenoid cystic carcinomas of major salivary glands.

PURPOSE: To report outcomes in patients treated with postoperative radiotherapy for nonadenoid cystic carcinomas of the major salivary glands. MATERIALS AND METHODS: From 1998-2011, 37 patients with nonadenoid cystic carcinomas of the major salivary gland underwent postoperative radiotherapy. The median radiation dose was 60 Gy (range, 45-70 Gy). TNM distribution included T1-2 (n=16, 44%), T3-T4 (n=21, 56%), N0 (n=19, 51%), and N+ (n=18, 49%). Histologies included adenocarcinoma (n=13, 35%), squamous cell carcinoma (n=8, 22%), mucoepidermoid carcinoma (n=8, 22%), and other (n=8, 21%). Median follow-up was 4.7 years for all patients (range, 0.3-14.1 years) and 5.0 years for living patients (range, 1.2-12.2 years). RESULTS: Five-year local-regional control, overall survival (OS), and cancer-specific survival (CSS) were 97%, 76%, and 84%. On univariate analysis, OS was significantly worse for patients ≥65 years old (p=0.04). CSS was significantly worse for positive perineural invasion (p=0.02), extraparenchymal extension (p=0.04), and in patients who received no chemotherapy (p=0.02). Doses >60 Gy was significantly worse for OS (p=0.003) and CSS (p=0.003), although these patients had higher TNM (>T2, p=0.01) and trended towards a higher rate of extraparenchymal extension (p=0.08). Four patients (11%) developed ≥grade 2 toxicities; 3 patients developed early toxicities and one patient developed late toxicities. CONCLUSIONS: Radiotherapy for salivary gland tumors provides excellent local-regional control when combined with surgery. Distant metastasis is the predominant pattern of failure, although chemotherapy seemed to improve cancer-specific survival.