Prediction of skin dose in low-kV intraoperative radiotherapy using machine learning models trained on results of in vivo dosimetry

Skin dose-volume predictors of moderate-severe late side effects after whole breast radiotherapy

BACKGROUND

Toxicity after whole breast Radiotherapy is a relevant issue, impacting the quality-of-life of a not negligible number of patients. We aimed to develop a Normal Tissue Complication Probability (NTCP) model predicting late toxicities by combining dosimetric parameters of the breast dermis and clinical factors.

METHODS

The skin structure was defined as the outer CT body contour's 5 mm inner isotropic expansion. It was retrospectively segmented on a large mono-institutional cohort of early-stage breast cancer patients enrolled between 2009 and 2017 (n = 1066). Patients were treated with tangential-field RT, delivering 40 Gy in 15 fractions to the whole breast. Toxicity was reported during Follow-Up (FU) using SOMA/LENT scoring. The study endpoint was moderate-severe late side effects consisting of Fibrosis-Atrophy-Telangiectasia-Pain (FATP G ≥ 2) developed within 42 months after RT completion. A machine learning pipeline was designed with a logistic model combining clinical factors and absolute skin DVH (cc) parameters as output.

RESULTS

The FATP G2 + rate was 3.8 %, with 40/1066 patients experiencing side effects. After the preprocessing of variables, a cross-validation was applied to define the best-performing model. We selected a 4-variable model with Post-Surgery Cosmetic alterations (Odds Ratio, OR = 7.3), Aromatase Inhibitors (as a protective factor with OR = 0.45), V20 Gy (50 % of the prescribed dose, OR = 1.02), and V42 Gy (105 %, OR = 1.09). Factors were also converted into an adjusted V20Gy.

CONCLUSIONS

The association between late reactions and skin DVH when delivering 40 Gy/15 fr was quantified, suggesting an independent role of V20 and V42. Few clinical factors heavily modulate the risk.

In vivodosimetry in cancer patients undergoing intraoperative radiation therapy

In vivodosimetry (IVD) is an important tool in external beam radiotherapy (EBRT) to detect major errors by assessing differences between expected and delivered dose and to record the received dose by individual patients. Also, in intraoperative radiation therapy (IORT), IVD is highly relevant to register the delivered dose. This is especially relevant in low-risk breast cancer patients since a high dose of IORT is delivered in a single fraction. In contrast to EBRT, online treatment planning based on intraoperative imaging is only under development for IORT. Up to date, two commercial treatment planning systems proposed intraoperative ultrasound or in-room cone-beam CT for real-time IORT planning. This makes IVD even more important because of the possibility for real-time treatment adaptation. Here, we summarize recent developments and applications of IVD methods for IORT in clinical practice, highlighting important contributions and identifying specific challenges such as a treatment planning system for IORT. HDR brachytherapy as a delivery technique was not considered. We add IVD for ultrahigh dose rate (FLASH) radiotherapy that promises to improve the treatment efficacy, when compared to conventional radiotherapy by limiting the rate of toxicity while maintaining similar tumour control probabilities. To date, FLASH IORT is not yet in clinical use.

Gold-nanoparticle-enriched breast tissue in breast cancer treatment using the INTRABEAM® system: a Monte Carlo study

Using a 50-kV INTRABEAM^® system after breast-conserving surgery, breast skin injury and long treatment time remain the challenging problems when large-size spherical applicators are used. This study has aimed to address these problems using gold (Au) nanoparticles (NPs). For this, surface and isotropic doses were measured using a Gafchromic EBT3 film and a water phantom. The particle propagation code EGSnrc/Epp was used to score the corresponding doses using a geometry similar to that used in the measurements. The simulation was validated using a gamma index of 2%/2 mm acceptance criterion in the gamma analysis. After validation Au-NP-enriched breast tissue was simulated to quantify any breast skin dose reduction and shortening of treatment time. It turned out that the gamma value deduced for validation of the simulation was in an acceptable range (i.e., less than one). For 20 mg-Au/g-breast tissue, the calculated Dose Enhancement Ratio (DER) of the breast skin was 0.412 and 0.414 using applicators with diameters of 1.5 cm and 5 cm, respectively. The corresponding treatment times were shortened by 72.22% and 72.30% at 20 mg-Au/g-breast tissue concentration, respectively. It is concluded that Au-NP-enriched breast tissue shows significant advantages, such as reducing the radiation dose received by the breast skin as well as shortening the treatment time. Additionally, the DERs were not significantly dependent on the size of the applicators.

Combining computed tomography and biologically effective dose in radiomics and deep learning improves prediction of tumor response to robotic lung stereotactic body radiation therapy

PURPOSE

The aim of this study is to improve the performance of machine learning (ML) models in predicting response of non-small cell lung cancer (NSCLC) to stereotactic body radiation therapy (SBRT) by integrating image features from pre-treatment computed tomography (CT) with features from the biologically effective dose (BED) distribution.

MATERIALS AND METHODS

Image features, consisting of crafted radiomic features or machine-learned features extracted using a convolutional neural network, were calculated from pre-treatment CT data and from dose distributions converted into BED for 80 NSCLC lesions over 76 patients treated with robotic guided SBRT. ML models using different combinations of features were trained to predict complete or partial response according to response criteria in solid tumors, including radiomics CT (Rad_CT ), radiomics CT and BED (Rad_CT,BED ), deep learning (DL) CT (DL_CT ), and DL CT and BED (DL_CT,BED ). Training of ML included feature selection by neighborhood component analysis followed by ensemble ML using robust boosting. A model was considered as acceptable when the sum of average sensitivity and specificity on test data in repeated cross validations was at least 1.5.

RESULTS

Complete or partial response occurred in 58 out of 80 lesions. The best models to predict the tumor response were those using BED variables, achieving significantly better area under curve (AUC) and accuracy than those using only features from CT, including a Rad_CT,BED model using three radiomic features from BED, which scored an accuracy of 0.799 (95% confidence intervals (0.75-0.85)) and AUC of 0.773 (0.688-0.846), and a DL_CT,BED model also using three variables with an accuracy of 0.798 (0.649-0.829) and AUC of 0.812 (0.755-0.867).

CONCLUSION

According to our results, the inclusion of BED features improves the response prediction of ML models for lung cancer patients undergoing SBRT, regardless of the use of radiomic or DL features.

Artificial Intelligence in Digital Pathology: What Is the Future? Part 1: From the Digital Slide Onwards

This commentary aims to address the field of Artificial intelligence (AI) in Digital Pathology (DP) both in terms of the global situation and research perspectives. It has four polarities. First, it revisits the evolutions of digital pathology with particular care to the two fields of the digital cytology and the digital histology. Second, it illustrates the main fields in the employment of AI in DP. Third, it looks at the future directions of the research challenges from both a clinical and technological point of view. Fourth, it discusses the transversal problems among these challenges and implications and introduces the immediate work to implement.

Artificial Intelligence and the Medical Physicist: Welcome to the Machine

Artificial intelligence (AI) is a branch of computer science dedicated to giving machines or computers the ability to perform human-like cognitive functions, such as learning, problem-solving, and decision making. Since it is showing superior performance than well-trained human beings in many areas, such as image classification, object detection, speech recognition, and decision-making, AI is expected to change profoundly every area of science, including healthcare and the clinical application of physics to healthcare, referred to as medical physics. As a result, the Italian Association of Medical Physics (AIFM) has created the “AI for Medical Physics” (AI4MP) group with the aims of coordinating the efforts, facilitating the communication, and sharing of the knowledge on AI of the medical physicists (MPs) in Italy. The purpose of this review is to summarize the main applications of AI in medical physics, describe the skills of the MPs in research and clinical applications of AI, and define the major challenges of AI in healthcare.