Automatic dose prediction using deep learning and plan optimization with finite-element control for intensity modulated radiation therapy.

BACKGROUND: Automatic solutions for generating radiotherapy treatment plans using deep learning (DL) have been investigated by mimicking the voxel's dose. However, plan optimization using voxel-dose features has not been extensively studied. PURPOSE: This study aims to investigate the efficiency of a direct optimization strategy with finite elements (FEs) after DL dose prediction for automatic intensity-modulated radiation therapy (IMRT) treatment planning. METHODS: A double-UNet DL model was adapted for 220 cervical cancer patients (200 for training and 20 for testing), who underwent IMRT between 2016 and 2020 at our clinic. The model inputs were computed tomography (CT) slices, organs at risk (OARs), and planning target volumes (PTVs), and the outputs were dose distributions of uniformly generated high-dose region-controlled plans. The FEs were discretized into equal intervals of the dose prediction value within the [OARs avoid PTV(O-P)] and [body avoids OARs & PTV(B-OP)] regions in the test cohort and used to define the objectives for IMRT plan optimization. The plans were optimized using a two-step process. In the beginning, the plans of two extra cases with and without low-dose region control were compared to pursue robust and optimal dose adjustment degree pattern of FEs. In the first step, the mean dose of O-P FEs were constrained to differing degrees according to the pattern. The further the FEs from the PTV, the tighter the constraints. In the second step, the mean dose of O-P FEs from first step were constrained again but weakly and the dose of the B-OP FEs from dose prediction and PTV were tightly regulated. The dosimetric parameters of the OARs and PTV were evaluated and compared using an interstep approach. In another 10 cases, the plans optimized via the aforementioned steps (method 1) were compared with those directly generated by the double-UNet dose prediction model trained by low and high region-controlled plans (method 2). RESULTS: The mean differences in dose metrics between the UNet-predicted dose and the clinical plans were: 0.47 Gy for bladder D50% ; 0.62 Gy for rectum D50% ; 0% for small intestine V30Gy ; 1% for small intestine V40Gy ; 4% for left femoral head V30Gy ; and 6% for right femoral head V30Gy . The reductions in mean dose (p < 0.001) after FE-based optimization were: 4.0, 1.9, 2.8, 5.9, and 5.7 Gy for the bladder, rectum, small intestine, left femoral head, and right femoral head, respectively, with flat PTV homogeneity and conformity. Method 1 plans produced lower mean doses than those of method 2 for the bladder (0.7 Gy), rectum (1.0 Gy), and small intestine (0.6 Gy), while maintaining  PTV homogeneity and conformity. CONCLUSION: FE-based direct optimization produced lower OAR doses and adequate PTV doses after DL prediction. This solution offers rapid and automatic plan optimization without manual adjustment, particularly in low-dose regions.

Efficacy and Failure Patterns of Early SBRT to the Primary Tumor in Advanced EGFR-Mutation-Positive Lung Cancer with EFGR-TKI Treatment: A Prospective, Single Arm, Phase II Study.

Early stereotactic body radiation therapy (SBRT) to the primary tumor combined with epidermal growth factor receptor tyrosine kinase inhibitor (EFGR-TKI) treatment may increase progression-free survival (PFS) by delaying resistance in patients with advanced EGFR-mutant non-small cell lung cancer (NSCLC). In this prospective, single arm, phase II study, patients with advanced NSCLC were treated with EGFR-TKI (icotinib 125 mg tid or gefitinib 250 mg qd) for one month followed by SBRT (40-60 Gy/5-8 F/5-10 d) to the primary tumor with concurrent EGFR-TKI until disease progression. The primary endpoint was PFS and the patterns of failure. Overall survival (OS) and adverse effects (AEs) were secondary endpoints. Overall, 41 advanced NSCLC patients with EGFR mutations received treatment with 24.42 months of median follow-up time. On average, SBRT was initiated 1.49 months after EGFR-TKI administration. Tumors were found to have an average shrinkage rate of 42.50%. Median PFS was 15.23 months (95% CI 13.10-17.36), while median OS was 27.57 months (95% CI 23.05-32.09). Thirty-three patients were found to have disease progression, of which new site failure (NF) (22 patients, 66.66%) was the most common pattern, followed by original site failure (OF) (7 patients, 21.21%) and simultaneous OF/NF (ONF) (4 patients, 12.12%). There were no Aes equal to or greater than grade 3, with the most frequent AE being radiation pneumonitis. Therefore, administering therapy targeted at the primary tumor using early SBRT after EGFR-TKI initiation is a new potentially safe and effective approach to treat EGFR-mutant advanced NSCLC.

Increased ATG5 Expression Predicts Poor Prognosis and Promotes EMT in Cervical Carcinoma.

Cervical cancer has the second-highest incidence and mortality of female malignancy. The major causes of mortality in patients with cervical cancer are invasion and metastasis. The epithelial-mesenchymal transition (EMT) process plays a major role in the acquisition of metastatic potential and motility. Autophagy-related genes (ARGs) are implicated in the EMT process, and autophagy exerts a dual function in EMT management at different phases of tumor progression. However, the role of specific ARGs during the EMT process has not yet been reported in cervical cancer. Based on the data from the Cancer Genome Atlas (TCGA) cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC) sequencing database, we performed the prognosis analysis for those ARGs obtained from the Human Autophagy database. ATG5 was identified as the only important harmful marker influencing survival of cervical cancer patients by univariate Cox regression (HR 1.7; 95% CI: 1.0-2.8, p = 0.047), and the 5-years survival rate for the high- and low-ATG5 expression groups was 0.486 (0.375-0.631) and 0.782 (0.708-0.863), respectively. TCGA CESC methylation data showed that eight methylation sites of ATG5 could also be significantly associated with the overall survival (OS) of cervical cancer patients. Single-sample gene-set enrichment and gene functional enrichment results showed that ATG5 was correlated with some cancer-related pathways, such as phagocytosis-related genes, endocytosis-related genes, immune-related genes, EMT score, and some EMT signature-related genes. Next, cell migration and invasion assay and Western blot were applied to detect the function of ATG5 in EMT of cervical cancer. In cervical cancer cells, ATG5 knockdown resulted in attenuation of migration and invasion. The functional study showed that knockdown of ATG5 could reverse EMT process by P-ERK, P-NFκBp65, P-mTOR pathways, and so on. In conclusion, the present study implies that ATG5 was a major contributor to EMT regulation and poor prognosis in cervical cancer.

Corrigendum: Increased ATG5 Expression Predicts Poor Prognosis and Promotes EMT in Cervical Carcinoma.

[This corrects the article DOI: 10.3389/fcell.2021.757184.].