Machine learning-based normal tissue complication probability model for predicting albumin-bilirubin (ALBI) grade increase in hepatocellular carcinoma patients.

PURPOSE: The aim of this study was to develop a normal tissue complication probability model using a machine learning approach (ML-based NTCP) to predict the risk of radiation-induced liver disease in hepatocellular carcinoma (HCC) patients. MATERIALS AND METHODS: The study population included 201 HCC patients treated with radiotherapy. The patients' medical records were retrospectively reviewed to obtain the clinical and radiotherapy data. Toxicity was defined by albumin-bilirubin (ALBI) grade increase. The normal liver dose-volume histogram was reduced to mean liver dose (MLD) based on the fraction size-adjusted equivalent uniform dose (2 Gy/fraction and α/ß = 2). Three types of ML-based classification models were used, a penalized logistic regression (PLR), random forest (RF), and gradient-boosted tree (GBT) model. Model performance was compared using the area under the receiver operating characteristic curve (AUROC). Internal validation was performed by 5-fold cross validation and external validation was done in 44 new patients. RESULTS: Liver toxicity occurred in 87 patients (43.1%). The best individual model was the GBT model using baseline liver function, liver volume, and MLD as inputs and the best overall model was an ensemble of the PLR and GBT models. An AUROC of 0.82 with a standard deviation of 0.06 was achieved for the internal validation. An AUROC of 0.78 with a standard deviation of 0.03 was achieved for the external validation. The behaviors of the best GBT model were also in good agreement with the domain knowledge on NTCP. CONCLUSION: We propose the methodology to develop an ML-based NTCP model to estimate the risk of ALBI grade increase.

Evaluation and external validation of minor lymphatic pelvic pathway for distant metastases in cervical cancer patients treated with concurrent chemoradiotherapy.

To evaluate and validate the minor lymphatic pathway for distant metastases in cervical cancer. This is a retrospective cohort of cervical cancer patients underwent curative concurrent chemoradiotherapy. We used original dataset from 1 university hospital and validation dataset from 3 university hospitals. Lymphadenopathy status in CT imaging was reviewed by radiologist in either the obturator and external iliac nodes (major pathway) or the internal iliac and presacral nodes (minor pathway). We then used Cox regression to adjust for all potential confounders, including paraaortic nodes, T stage, histology, age, total treatment time, total number of nodes, total short axis of nodes. 397 and 384 patients were in the original and validation datasets (median follow-up period, 59.5 month's). The minor pathway was independent prognostic factor in multivariable analysis [HR=2.64; 95%CI=1.07-6.55; P = 0.036] and [HR=14.84; 95%CI=3.15-70.01; P= 0.001] in original and validation datasets, respectively. Whereas, the major pathway was statistically non-significant. Further validation showed that the minor pathway had the highest HR for distant metastases with both the EMBRACE (HR=6.05; 95% CI=1.30-28.08; P = 0.022) and the FIGO 2018 (HR=7.43; 95% CI=2.94-18.78; P<0.001) in the original dataset. A similar result was found with the validation dataset: EMBRACE, HR=30.91; 95% CI=2.78-343.62; P = 0.005; and FIGO 2018, HR=42.41; 95% CI=8.83-203.60; P<0.001.This is the first clinical study to validate that the minor lymphatic pathway was predominantly associated with distant metastases in cervical cancer. This finding should be validated in larger cohort to further integrate in standard staging for prediction of distant metastases.

Development and validation of a prognostic prediction model including the minor lymphatic pathway for distant metastases in cervical cancer patients.

To develop and validate a prognostic model, including the minor lymphatic pathway (internal iliac and presacral nodes). STUDY DESIGN: Retrospective cohort. PARTICIPANTS: Locally advanced cervical cancer underwent concurrent chemoradiotherapy. SAMPLE SIZE: 397 and 384 patients in the development and validation data set. PREDICTORS: Our new nodal staging system with the minor lymphatic pathway. OUTCOME: Distant metastases. STATISTICAL ANALYSIS: Cox regression; net reclassification improvement (NRI) and decision curve analysis (DCA). Our new nodal system was the strongest predictor. The predictors in the final model were new nodal system, tumor stage, adenocarcinoma, initial hemoglobin, tumor size and age. The nodal system and the pretreatment model had concordance indices of 0.661 and 0.708, respectively, with good calibration curves. Compared to the OUTBACK eligibility criteria, the nodal system showed NRI for both cases (22%) and controls (16%). The pretreatment model showed NRI for cases (31%) and controls (18%). DCA in both models showed threshold probability of 15% and 12%, respectively, when compared with 24% in OUTBACK eligibility criteria. Our new nodal staging system and the pretreatment model could differentiate between high-risk and low-risk patients, thus facilitating decisions to provide more aggressive treatment to prevent distant metastases.

Advanced radiotherapy technique in hepatocellular carcinoma with portal vein thrombosis: Feasibility and clinical outcomes.

BACKGROUND: In Thailand, individuals with hepatocellular carcinoma (HCC) who develop portal vein tumor thrombosis (PVTT) have a restricted treatment option because to the extent of the disease, poor underlying liver function, and non-coverage of immuno/targeted therapy. Radiotherapy (RT) plays an increasingly important function in these patients. To investigate the feasibility, efficacy, and adverse event rates, we performed a retrospective analysis of patients with HCC with PVTT who underwent 3-dimensional conformal radiation (3DCRT), intensity-modulated radiation (IMRT), volumetric-modulated radiotherapy (VMAT), and stereotactic body radiotherapy (SBRT) in a single-institution. OBJECTIVES: To examine clinical results in terms of overall survival (OS), local control (LC), response of primary tumor and PVTT, hepatic and gastrointestinal adverse reaction, and prognosis variables for OS and LC. MATERIALS AND METHODS: Between July 2007 and August 2019, non-metastatic HCC with PVTT patients treated with RT were retrospectively reviewed and evaluated. RESULTS: The analysis included data from 160 patients. The mean age of the patients was 60.8 years ((95% CI 58.2-62.0). The median diameter of the tumor was 7.7 cm (range: 1-24.5). 85 (54.5%) individuals had PVTT in the main or first branch. At 1.8-10 Gy per fraction, the mean biologically effective dose (BED) as α/ß ratio of 10 was 49.6 (95% CI 46.7-52.5) Gy10. The median survival time was 8.3 (95% CI 6.1-10.3) months. Survival rates at one and two years were 39.6% and 17.1%, respectively. Estimated incidence of local failure using competing risk analysis were 24% and 60% at 1 and 2 years, respectively. The overall response rate was 74%, with an 18.5 percent complete response rate. In multivariate analysis, tumor size, overall response, and radiation dose were all significant prognostic variables for OS. Hepatic unfavorable events of grade 3 and 4 were for 14.1% of the total. There was no occurrences of grade 3-4 gastrointestinal toxicity, either acute or late. Additionally, there were no treatment-related mortality. CONCLUSIONS: Advanced RT is regarded as a safe and effective therapeutic option for HCC with PVTT. Overall survival was clearly related to tumor size, radiation dose, and tumor/PVTT response. Individuals with BED 56 Gy10 had significantly better overall survival than patients with BED 56 Gy10. A prospective randomized trial is required to validate these outcomes in order to corroborate these findings.

Assessment of the confidence interval in the multivariable normal tissue complication probability model for predicting radiation-induced liver disease in primary liver cancer.

We developed a confidence interval-(CI) assessing model in multivariable normal tissue complication probability (NTCP) modeling for predicting radiation-induced liver disease (RILD) in primary liver cancer patients using clinical and dosimetric data. Both the mean NTCP and difference in the mean NTCP (ΔNTCP) between two treatment plans of different radiotherapy modalities were further evaluated and their CIs were assessed. Clinical data were retrospectively reviewed in 322 patients with hepatocellular carcinoma (n = 215) and intrahepatic cholangiocarcinoma (n = 107) treated with photon therapy. Dose-volume histograms of normal liver were reduced to mean liver dose (MLD) based on the fraction size-adjusted equivalent uniform dose. The most predictive variables were used to build the model based on multivariable logistic regression analysis with bootstrapping. Internal validation was performed using the cross-validation leave-one-out method. Both the mean NTCP and the mean ΔNTCP with 95% CIs were calculated from computationally generated multivariate random sets of NTCP model parameters using variance-covariance matrix information. RILD occurred in 108/322 patients (33.5%). The NTCP model with three clinical and one dosimetric parameter (tumor type, Child-Pugh class, hepatitis infection status and MLD) was most predictive, with an area under the receiver operative characteristics curve (AUC) of 0.79 (95% CI 0.74-0.84). In eight clinical subgroups based on the three clinical parameters, both the mean NTCP and the mean ΔNTCP with 95% CIs were able to be estimated computationally. The multivariable NTCP model with the assessment of 95% CIs has potential to improve the reliability of the NTCP model-based approach to select the appropriate radiotherapy modality for each patient.

Flattening filter free stereotactic body radiation therapy for lung tumors: outcomes and predictive factors.

BACKGROUND: Stereotactic body radiation therapy (SBRT) using flattening filter free (FFF) has been commonly used, however, its outcomes and predictive factors in lung tumors are limiting. Thus, we aim to assess the clinical outcomes of this approach and identify factors associated with outcomes in patients with early stage non-small cell lung cancer (NSCLC) and oligometastatic/oligoprogressive lung tumor (OLT). METHODS: Patients who underwent lung SBRT with FFF were retrospectively reviewed. All patients were delivered using volumetric modulated arc therapy (VMAT) technique. The primary outcome was local control (LC). The secondary outcomes were overall survival (OS) and toxicities. We assessed the association between LC and various factors in OLT. RESULTS: From February 2014 to July 2019, ninety-four patients and 129 lesions with median follow-up time of 30 months were included in the analysis. Twenty-six patients with 26 lesions were early NSCLC, while 68 patients with 103 lesions were OLT, 41.7% of which were from colorectal cancers (CRC) and 18.5% were from primary lung cancers. Two-year LC was 88.9% and 85.7% for early NSCLC and OLT, respectively. Two-year OS was significantly higher for early NSCLC than OLT (83.3% vs. 68.7%, P=0.035). In the multivariate analysis for OLT, CRC origin (hazard ratio, HR 10.59, 95% CI: 2.29-48.95, P=0.003) and gross tumor volume (GTV) mean BED10 ≤147 Gy (HR 5.16, 95% CI: 1.13-23.59, P=0.034) were significantly associated with higher local failure (LF). Most of the acute grade 1-2 toxicities were radiation pneumonitis (26.5%). No grade 3-5 event was observed. CONCLUSIONS: This study confirmed the clinical efficacy and safety of lung SBRT using FFF-technique. Our findings support the role of using a high BED10 regimen to achieve good LC for OLT and the potential role for dose escalation for primary CRC.

The normal tissue complication probability model-based approach considering uncertainties for the selective use of radiation modality in primary liver cancer patients.

PURPOSE: To predict the probability of radiation-induced liver toxicity (RILT) and implement the normal tissue complication probability (NTCP) model-based approach considering confidence intervals (CIs) to select patients for new treatment techniques, such as proton beam therapy, based on a certain NTCP reduction (ΔNTCP) threshold for primary liver cancer patients. METHODS AND MATERIALS: Common Toxicity Criteria for Adverse Events (CTCAE) grade ≥2 RILT was scored. The Lyman NTCP models predicting the probability of CTCAE grade ≥2 RILT as a function of the fraction-size adjusted mean liver dose (MLD), using reference fraction size = 2 Gy/fraction and α/ß ratio = 2 Gy, were fitted using the maximum likelihood method. At certain combinations of MLDs, ΔNTCP with a CI was evaluated by the delta method. RESULTS: Of the 239 patients, the incidence of CTCAE grade ≥2 RILT was 55% (46% in the Child-Pugh (CP)-A vs. 81% in the CP-B/C, p < 0.001). Among 180 CP-A patients, 40% who had viral hepatitis infections experienced toxicity vs. 32% in the nonhepatitis subgroup. The MLD was 18 Gy in the toxicity group vs. 16.1 Gy in the nontoxicity group (p = 0.002). The estimated NTCP model parameters specific to the patient subgroups and the ΔNTCP with CI assuming a particular CP classification and viral hepatitis infection status were considerably different which possible changed treatment decision. CONCLUSIONS: Patients with CP-A and viral hepatitis infection or CP-B/C cirrhosis had greater susceptibility to CTCAE grade ≥2 RILT. The estimated NTCP and ΔNTCP for individual patients along with a consideration of uncertainties improve the reliability of the NTCP model-based approach.