This study reviewed and meta-analyzed evidence on radiomics-based hybrid models for predicting radiation pneumonitis (RP). These models are crucial for improving thoracic radiotherapy plans and mitigating RP, a common complication of thoracic radiotherapy. We examined and compared the RP prediction models developed in these studies with the radiomics features employed in RP models. We systematically searched Google Scholar, Embase, PubMed, and MEDLINE for studies published up to April 19, 2024. Sixteen studies met the inclusion criteria. We compared the RP prediction models developed in these studies and the radiomics features employed. Radiomics, as a single-factor evaluation, achieved an area under the receiver operating characteristic curve (AUROC) of 0.73, accuracy of 0.69, sensitivity of 0.64, and specificity of 0.74. Dosiomics achieved an AUROC of 0.70. Clinical and dosimetric factors showed lower performance, with AUROCs of 0.59 and 0.58. Combining clinical and radiomic factors yielded an AUROC of 0.78, while combining dosiomic and radiomics factors produced an AUROC of 0.81. Triple combinations, including clinical, dosimetric, and radiomics factors, achieved an AUROC of 0.81. The study identifies key radiomics features, such as the Gray Level Co-occurrence Matrix (GLCM) and Gray Level Size Zone Matrix (GLSZM), which enhance the predictive accuracy of RP models. Radiomics-based hybrid models are highly effective in predicting RP. These models, combining traditional predictive factors with radiomic features, particularly GLCM and GLSZM, offer a clinically feasible approach for identifying patients at higher RP risk. This approach enhances clinical outcomes and improves patient quality of life. The protocol of this study was registered on PROSPERO (CRD42023426565).

To reduce radiation doses to the tongue, a patient-specific semi-customized tongue displacement device (SCTDD) was developed using a 3D printer (3DP) for head and neck (H&N) radiation therapy (RT). Dosimetric characteristics of the SCTDD were compared with those of a standard mouthpiece (SMP). The SCTDD consists of three parts: a mouthpiece, connector with an immobilization mask, and tongue displacer, which can displace the tongue to the contralateral side of the planning target volume. Semi-customization was enabled by changing the thickness and length of the SCTDD. The instrument was printed using a 3DP with a biocompatible material. With the SCTDD and SMP, two sets of planning computed tomography (CT) and tomotherapy plans were obtained for seven H&N cancer patients. Dosimetric and geometric characteristics were compared. Using the SCTDD, the tongue was effectively displaced from the planning target volume without significant tongue volume change compared to the SMP. The median tongue dose was significantly reduced (29.6 Gy vs. 34.3 Gy). The volumes of the tongue receiving a dose of 15 Gy, 30 Gy, 35 Gy, 45 Gy, and 60 Gy were significantly lower than using the SMP. The SCTDD significantly decreased the radiation dose to the tongue compared to the SMP, which may potentially reduce RT-related tongue toxicity.

Purpose: The European Society of Radiation & Oncology and Advisory Committee on Radiation Oncology Practice (ESTRO-ACROP) presented new guidelines for clinical target volume (CTV) delineation in post-mastectomy radiation therapy (PMRT) after implant-based immediate breast reconstruction (IBR-i). This study evaluated the dosimetric characteristics, dosimetric accuracy, and delivery accuracy of these guidelines in volumetric modulated arc therapy (VMAT). Methods and Materials: This retrospective study included 15 patients with left breast cancer who underwent mastectomy with tissue expander placement followed by PMRT. An experienced radiation oncologist delineated the CTV twice on the same image datasets based on the ESTRO-ACROP (EA-TVD) and conventional target volume delineation (C-TVD) guidelines. All VMAT plans, which used a double partial arc, were generated using six MV photons. Clinically relevant dose-volume parameters for organs at risk were compared. Dosimetric accuracy of the treatment plans and delivery accuracy were assessed. Results: Target volume of EA-TVD was significantly smaller than that of C-TVD. Although no statistically significant difference was noted in the target coverage between the two VMAT plans, EA-TVD VMAT significantly reduced the mean heart dose (3.99 ± 1.02 vs. 5.84 ± 1.78 Gy, p = 0.000), the maximum left anterior descending coronary artery (LAD) dose (9.43 ± 3.04 vs. 13.97 ± 6.04 Gy, p = 0.026), and the mean LAD dose (4.52 ± 1.31 vs. 6.35 ± 2.79 Gy, p = 0.028) compared with C-TVD VMAT. No significant difference was observed with respect to the total monitor units, plan complexity, and delivery quality assurance. Conclusions: This is the first study to show significant dose reduction for the normal heart and LAD tissue offered by the EA-TVD, while maintaining dosimetric and delivery accuracy, in PMRT after IBR-i in VMAT for left-sided breast cancer patients.

To develop and test a multivariable normal tissue complication probability (NTCP) model predicting lymphedema in patients with breast cancer receiving radiation therapy. We retrospectively reviewed 1345 patients with breast cancer who received radiation therapy from 2 independent institutions. The patients were divided into a training cohort (institution A, n = 368, all treated with 3-dimensional conformal external beam radiation therapy [RT] with 2 Gy/fraction) and an external validation cohort (institution B, n = 977, treated either with 3-dimensional conformal external beam RT or with volumetric modulated RT and either with 1.8-2.0 Gy/fraction or with 2.67 Gy/fraction). Axillary-lateral thoracic vessel juncture (ALTJ) was delineated. The multivariable model was generated using dosimetric and clinical parameters. The performance of the model was comprehensively validated internally and externally. During a median follow-up of 78.7 months for the entire cohort, 97 patients (7.2%) developed lymphedema. The multivariable model that took into account the number of lymph nodes dissected, as well as the volume of the ALTJ receiving a dose ≥35 Gy equivalent doses in 2-Gy fractions (ALTJ V35), showed good agreement between predicted and observed results for both internal and external validation (Hosmer-Lemeshow P value > .05). The area under the receiver operating characteristic curve (AUC) and negative log-likelihood values for the multivariable NTCP model were 0.89 and 0.19 in internal validation and 0.83 and 0.19 in external validation. In addition, the multivariable model performance was acceptable for hypofractionated regimens (AUC 0.70) and volumetric modulated arc therapy (AUC 0.69). The number of lymph nodes dissected and ALTJ V35 were found to be the most important factors influencing lymphedema after radiation therapy. We first developed and validated the multivariable NTCP model for the lymphedema incidence in patients with breast cancer after radiation therapy. The multivariable NTCP model showed excellent performance and robustness in predicting lymphedema in both internal and completely independent external validations. The multivariable model for lymphedema prediction was robust and reliable for different treatment modalities and fractionation regimens.

Objective evaluation of radiation dermatitis is important for analysing the correlation between the severity of radiation dermatitis and dose distribution in clinical practice and for reliable reporting in clinical trials. We developed a novel radiation dermatitis segmentation system based on convolutional neural networks (CNNs) to consistently evaluate radiation dermatitis. The radiation dermatitis segmentation system is designed to segment the radiation dermatitis occurrence area using skin photographs and skin-dose distribution. A CNN architecture with a dilated convolution layer and skip connection was designed to estimate the radiation dermatitis area. Seventy-three skin photographs obtained from patients undergoing radiotherapy were collected for training and testing. The ground truth of radiation dermatitis segmentation is manually delineated from the skin photograph by an experienced radiation oncologist and medical physicist. We converted the skin photographs to RGB (red-green-blue) and CIELAB (lightness (L^∗), red-green (a^∗) and blue-yellow (b^∗)) colour information and trained the network to segment faint and severe radiation dermatitis using three different input combinations: RGB, RGB + CIELAB (RGBLAB) and RGB + CIELAB + skin-dose distribution (RGBLAB_D). The proposed system was evaluated using the Dice similarity coefficient (DSC), sensitivity, specificity and normalised Matthews correlation coefficient (nMCC). A paired t-test was used to compare the results of different segmentation performances. Optimal data composition was observed in the network trained for radiation dermatitis segmentation using skin photographs and skin-dose distribution. The average DSC, sensitivity, specificity and nMCC values of RGBLAB_D were 0.62, 0.61, 0.91 and 0.77, respectively, in faint radiation dermatitis, and 0.69, 0.78, 0.96 and 0.83, respectively, in severe radiation dermatitis. Our study showed that CNN-based radiation dermatitis segmentation in skin photographs of patients undergoing radiotherapy can describe radiation dermatitis severity and pattern. Our study could aid in objectifying the radiation dermatitis grading and analysing the reliable correlation between dosimetric factors and the morphology of radiation dermatitis.