Brain and Brain Stem Necrosis After Reirradiation for Recurrent Childhood Primary Central Nervous System Tumors: A PENTEC Comprehensive Review.

PURPOSE: Reirradiation is increasingly used in children and adolescents/young adults (AYA) with recurrent primary central nervous system tumors. The Pediatric Normal Tissue Effects in the Clinic (PENTEC) reirradiation task force aimed to quantify risks of brain and brain stem necrosis after reirradiation. METHODS AND MATERIALS: A systematic literature search using the PubMed and Cochrane databases for peer-reviewed articles from 1975 to 2021 identified 92 studies on reirradiation for recurrent tumors in children/AYA. Seventeen studies representing 449 patients who reported brain and brain stem necrosis after reirradiation contained sufficient data for analysis. While all 17 studies described techniques and doses used for reirradiation, they lacked essential details on clinically significant dose-volume metrics necessary for dose-response modeling on late effects. We, therefore, estimated incidences of necrosis with an exact 95% CI and qualitatively described data. Results from multiple studies were pooled by taking the weighted average of the reported crude rates from individual studies. RESULTS: Treated cancers included ependymoma (n = 279 patients; 7 studies), medulloblastoma (n = 98 patients; 6 studies), any CNS tumors (n = 62 patients; 3 studies), and supratentorial high-grade gliomas (n = 10 patients; 1 study). The median interval between initial and reirradiation was 2.3 years (range, 1.2-4.75 years). The median cumulative prescription dose in equivalent dose in 2-Gy fractions (EQD22; assuming α/ß value = 2 Gy) was 103.8 Gy (range, 55.8-141.3 Gy). Among 449 reirradiated children/AYA, 22 (4.9%; 95% CI, 3.1%-7.3%) developed brain necrosis and 14 (3.1%; 95% CI, 1.7%-5.2%) developed brain stem necrosis with a weighted median follow-up of 1.6 years (range, 0.5-7.4 years). The median cumulative prescription EQD22 was 111.4 Gy (range, 55.8-141.3 Gy) for development of any necrosis, 107.7 Gy (range, 55.8-141.3 Gy) for brain necrosis, and 112.1 Gy (range, 100.2-117 Gy) for brain stem necrosis. The median latent period between reirradiation and the development of necrosis was 5.7 months (range, 4.3-24 months). Though there were more events among children/AYA undergoing hypofractionated versus conventionally fractionated reirradiation, the differences were not statistically significant (P = .46). CONCLUSIONS: Existing reports suggest that in children/AYA with recurrent brain tumors, reirradiation with a total EQD22 of about 112 Gy is associated with an approximate 5% to 7% incidence of brain/brain stem necrosis after a median follow-up of 1.6 years (with the initial course of radiation therapy being given with conventional prescription doses of ≤2 Gy per fraction and the second course with variable fractionations). We recommend a uniform approach for reporting dosimetric endpoints to derive robust predictive models of late toxicities following reirradiation.

Comparison of deep learning networks for fully automated head and neck tumor delineation on multi-centric PET/CT images.

OBJECTIVES: Deep learning-based auto-segmentation of head and neck cancer (HNC) tumors is expected to have better reproducibility than manual delineation. Positron emission tomography (PET) and computed tomography (CT) are commonly used in tumor segmentation. However, current methods still face challenges in handling whole-body scans where a manual selection of a bounding box may be required. Moreover, different institutions might still apply different guidelines for tumor delineation. This study aimed at exploring the auto-localization and segmentation of HNC tumors from entire PET/CT scans and investigating the transferability of trained baseline models to external real world cohorts. METHODS: We employed 2D Retina Unet to find HNC tumors from whole-body PET/CT and utilized a regular Unet to segment the union of the tumor and involved lymph nodes. In comparison, 2D/3D Retina Unets were also implemented to localize and segment the same target in an end-to-end manner. The segmentation performance was evaluated via Dice similarity coefficient (DSC) and Hausdorff distance 95th percentile (HD95). Delineated PET/CT scans from the HECKTOR challenge were used to train the baseline models by 5-fold cross-validation. Another 271 delineated PET/CTs from three different institutions (MAASTRO, CRO, BERLIN) were used for external testing. Finally, facility-specific transfer learning was applied to investigate the improvement of segmentation performance against baseline models. RESULTS: Encouraging localization results were observed, achieving a maximum omnidirectional tumor center difference lower than 6.8 cm for external testing. The three baseline models yielded similar averaged cross-validation (CV) results with a DSC in a range of 0.71-0.75, while the averaged CV HD95 was 8.6, 10.7 and 9.8 mm for the regular Unet, 2D and 3D Retina Unets, respectively. More than a 10% drop in DSC and a 40% increase in HD95 were observed if the baseline models were tested on the three external cohorts directly. After the facility-specific training, an improvement in external testing was observed for all models. The regular Unet had the best DSC (0.70) for the MAASTRO cohort, and the best HD95 (7.8 and 7.9 mm) in the MAASTRO and CRO cohorts. The 2D Retina Unet had the best DSC (0.76 and 0.67) for the CRO and BERLIN cohorts, and the best HD95 (12.4 mm) for the BERLIN cohort. CONCLUSION: The regular Unet outperformed the other two baseline models in CV and most external testing cohorts. Facility-specific transfer learning can potentially improve HNC segmentation performance for individual institutions, where the 2D Retina Unets could achieve comparable or even better results than the regular Unet.

In Doxorubicin-Adapted Hodgkin Lymphoma Cells, Acquiring Multidrug Resistance and Improved Immunosuppressive Abilities, Doxorubicin Activity Was Enhanced by Chloroquine and GW4869.

Classical Hodgkin lymphoma (cHL) is a highly curable disease (70-80%), even though long-term toxicities, drug resistance, and predicting clinical responses to therapy are major challenges in cHL treatment. To solve these problems, we characterized two cHL cell lines with acquired resistance to doxorubicin, KM-H2dx and HDLM-2dx (HRSdx), generated from KM-H2 and HDLM-2 cells, respectively. HRSdx cells developed cross-resistance to vinblastine, bendamustin, cisplatin, dacarbazine, gemcitabine, brentuximab vedotin (BV), and γ-radiation. Both HDLM-2 and HDLM-2dx cells had intrinsic resistance to BV but not to the drug MMAE. HDLM-2dx acquired cross-resistance to caelyx. HRSdx cells had in common decreased CD71, CD80, CD54, cyt-ROS, HLA-DR, DDR1, and CD44; increased Bcl-2, CD58, COX2, CD26, CCR5, and invasive capability; increased CCL5, TARC, PGE2, and TGF-ß; and the capability of hijacking monocytes. In HRSdx cells less sensitive to DNA damage and oxidative stress, the efflux drug transporters MDR1 and MRP1 were not up-regulated, and doxorubicin accumulated in the cytoplasm rather than in the nucleus. Both the autophagy inhibitor chloroquine and extracellular vesicle (EV) release inhibitor GW4869 enhanced doxorubicin activity and counteracted doxorubicin resistance. In conclusion, this study identifies common modulated antigens in HRSdx cells, the associated cross-resistance patterns, and new potential therapeutic options to enhance doxorubicin activity and overcome resistance.

Clinical Use of a Commercial Artificial Intelligence-Based Software for Autocontouring in Radiation Therapy: Geometric Performance and Dosimetric Impact.

PURPOSE: When autocontouring based on artificial intelligence (AI) is used in the radiotherapy (RT) workflow, the contours are reviewed and eventually adjusted by a radiation oncologist before an RT treatment plan is generated, with the purpose of improving dosimetry and reducing both interobserver variability and time for contouring. The purpose of this study was to evaluate the results of application of a commercial AI-based autocontouring for RT, assessing both geometric accuracies and the influence on optimized dose from automatically generated contours after review by human operator. MATERIALS AND METHODS: A commercial autocontouring system was applied to a retrospective database of 40 patients, of which 20 were treated with radiotherapy for prostate cancer (PCa) and 20 for head and neck cancer (HNC). Contours resulting from AI were compared against AI contours reviewed by human operator and human-only contours using Dice similarity coefficient (DSC), Hausdorff distance (HD), and relative volume difference (RVD). Dosimetric indices such as Dmean, D0.03cc, and normalized plan quality metrics were used to compare dose distributions from RT plans generated from structure sets contoured by humans assisted by AI against plans from manual contours. The reduction in contouring time obtained by using automated tools was also assessed. A Wilcoxon rank sum test was computed to assess the significance of differences. Interobserver variability of the comparison of manual vs. AI-assisted contours was also assessed among two radiation oncologists for PCa. RESULTS: For PCa, AI-assisted segmentation showed good agreement with expert radiation oncologist structures with average DSC among patients ≥ 0.7 for all structures, and minimal radiation oncology adjustment of structures (DSC of adjusted versus AI structures ≥ 0.91). For HNC, results of comparison between manual and AI contouring varied considerably e.g., 0.77 for oral cavity and 0.11-0.13 for brachial plexus, but again, adjustment was generally minimal (DSC of adjusted against AI contours 0.97 for oral cavity, 0.92-0.93 for brachial plexus). The difference in dose for the target and organs at risk were not statistically significant between human and AI-assisted, with the only exceptions of D0.03cc to the anal canal and Dmean to the brachial plexus. The observed average differences in plan quality for PCa and HNC cases were 8% and 6.7%, respectively. The dose parameter changes due to interobserver variability in PCa were small, with the exception of the anal canal, where large dose variations were observed. The reduction in time required for contouring was 72% for PCa and 84% for HNC. CONCLUSIONS: When an autocontouring system is used in combination with human review, the time of the RT workflow is significantly reduced without affecting dose distribution and plan quality.

Risk of Subsequent Neoplasms in Childhood Cancer Survivors After Radiation Therapy: A Comprehensive PENTEC Review.

PURPOSE: A Pediatric Normal Tissue Effects in the Clinic (PENTEC) analysis of published investigations of central nervous system (CNS) subsequent neoplasms (SNs), subsequent sarcomas, and subsequent lung cancers in childhood cancer survivors who received radiation therapy (RT) was performed to estimate the effect of RT dose on the risk of SNs and the modification of this risk by host and treatment factors. METHODS AND MATERIALS: A systematic literature review was performed to identify data published from 1975 to 2022 on SNs after prior RT in childhood cancer survivors. After abstract review, usable quantitative and qualitative data were extracted from 83 studies for CNS SNs, 118 for subsequent sarcomas, and 10 for lung SNs with 4 additional studies (3 for CNS SNs and 1 for lung SNs) later added. The incidences of SNs, RT dose, age, sex, primary cancer diagnosis, chemotherapy exposure, and latent time from primary diagnosis to SNs were extracted to assess the factors influencing risk for SNs. The excess relative ratio (ERR) for developing SNs as a function of dose was analyzed using inverse-variance weighted linear regression, and the ERR/Gy was estimated. Excess absolute risks were also calculated. RESULTS: The ERR/Gy for subsequent meningiomas was estimated at 0.44 (95% CI, 0.19-0.68); for malignant CNS neoplasms, 0.15 (95% CI, 0.11-0.18); for sarcomas, 0.045 (95% CI, 0.023-0.067); and for lung cancer, 0.068 (95% CI, 0.03-0.11). Younger age at time of primary diagnosis was associated with higher risk of subsequent meningioma and sarcoma, whereas no significant effect was observed for age at exposure for risk of malignant CNS neoplasm, and insufficient data were available regarding age for lung cancer. Females had a higher risk of subsequent meningioma (odds ratio, 1.46; 95% CI, 1.22-1.76; P < .0001) relative to males, whereas no statistically significant sex difference was seen in risk of malignant CNS neoplasms, sarcoma SNs, or lung SNs. There was an association between chemotherapy receipt (specifically alkylating agents and anthracyclines) and subsequent sarcoma risk, whereas there was no clear association between specific chemotherapeutic agents and risk of CNS SNs and lung SNs. CONCLUSIONS: This PENTEC systematic review shows a significant radiation dose-response relationship for CNS SNs, sarcomas, and lung SNs. Given the linear dose response, improved conformality around the target volume that limits the high dose volume might be a promising strategy for reducing the risk of SNs after RT. Other host- and treatment-related factors such as age and chemotherapy play a significant contributory role in the development of SNs and should be considered when estimating the risk of SNs after RT among childhood cancer survivors.

Physics Considerations for Evaluation of Dose for Dose-Response Models of Pediatric Late Effects From Radiation Therapy.

PURPOSE: We describe the methods used to estimate the accuracy of dosimetric data found in literature sources used to construct the Pediatric Normal Tissue Effects in the Clinic (PENTEC) dose-response models, summarize these findings of each organ-specific task force, describe some of the dosimetric challenges and the extent to which these efforts affected the final modeling results, and provide guidance on the interpretation of the dose-response results given the various dosimetric uncertainties. METHODS AND MATERIALS: Each of the PENTEC task force medical physicists reviewed all the journal articles used for dose-response modeling to identify, categorize, and quantify dosimetric uncertainties. These uncertainties fell into 6 broad categories. A uniform nomenclature was developed for describing the "dosimetric quality" of the articles used in the PENTEC reviews. Among the multidisciplinary experts in the PENTEC effort, the medical physicists were charged with the dosimetric evaluation, as they are most expert in this subject. RESULTS: The percentage dosimetric uncertainty was estimated for each late effect endpoint for all PENTEC organ reports. Twelve specific sources of dose uncertainty were identified related to the 6 broad categories. The most common reason for organ dose uncertainty was that prescribed dose rather than organ dose was reported. Percentage dose uncertainties ranged from 5% to 200%. Systematic uncertainties were used to correct the dose component of the models. Random uncertainties were also described in each report and in some cases used to modify dose axis error bars. In addition, the potential effects of dose binning were described. CONCLUSIONS: PENTEC reports are designed to provide guidance to radiation oncologists and treatment planners for organ dose constraints. It is critical that these dose constraint recommendations are as accurate as possible, acknowledging the large error bars for many. Achieving this accuracy is important as it enables clinicians to better balance target dose coverage with risk of late effects. Evidence-based dose constraints for pediatric patients have been lacking and, in this regard, PENTEC fills an important unmet need. One must be aware of the limitations of our recommendations, and that for some organ systems, large uncertainties exist in the dose-response model because of clinical endpoint uncertainty, dosimetric uncertainty, or both.

Regulatory Aspects of the Use of Artificial Intelligence Medical Software.

The rapidly evolving scenario of Artificial intelligence (AI) in medicine comes with new regulatory challenges, including certification, ownership, and control of data sharing, privacy protection, and accountability. The Medical Physicists (MPs) are traditionally responsible for ensuring the safety and quality of technology implementation in diagnostic and therapeutic settings. As such, they are also expected to contribute to the introduction of AI medical devices in routine clinical practice. Specifically, the MPs will play a stakeholder role for AI tools procurement, acceptance testing, commissioning, and quality assurance to confirm the claimed performances in relation to the medical device's intended use. Moreover, MPs who act as co-creators of such AI tools, will play a pivotal role in product requirements definition, data collection and annotation, clinical evaluation, support for regulatory pathways and marketing through scientific congresses and scientific publications. As AI software differs from the traditional (hardware) medical device that the MP is used to introduce in clinical settings, there is a need to acquire new competencies in the field of AI and its regulatory aspects. The purpose of this paper is to provide MPs with practical guidelines on regulatory aspects of AI medical devices, in the European and in the US landscape.

Predicting Local Failure after Partial Prostate Re-Irradiation Using a Dosiomic-Based Machine Learning Model.

The aim of this study is to predict local failure after partial prostate re-irradiation for the treatment of isolated locally recurrent prostate cancer by using a machine learning classifier based on radiomic features from pre-treatment computed tomography (CT), positron-emission tomography (PET) and biological effective dose distribution (BED) of the radiotherapy plan. The analysis was conducted on a monocentric dataset of 43 patients with evidence of isolated intraprostatic recurrence of prostate cancer after primary external beam radiotherapy. All patients received partial prostate re-irradiation delivered by volumetric modulated arc therapy. The gross tumor volume (GTV) of each patient was manually contoured from planning CT, choline-PET and dose maps. An ensemble machine learning pipeline including unbalanced data correction and feature selection was trained using the radiomic and dosiomic features as input for predicting occurrence of local failure. The model performance was assessed using sensitivity, specificity, accuracy and area under receiver operating characteristic curves of the score function in 10-fold cross validation repeated 100 times. Local failure was observed in 13 patients (30%), with a median time to recurrence of 36.7 months (range = 6.1-102.4 months). A four variables ensemble machine learning model resulted in accuracy of 0.62 and AUC 0.65. According to our results, a dosiomic machine learning classifier can predict local failure after partial prostate re-irradiation.

Applications of artificial intelligence in stereotactic body radiation therapy.

This topical review focuses on the applications of artificial intelligence (AI) tools to stereotactic body radiation therapy (SBRT). The high dose per fraction and the limited number of fractions in SBRT require stricter accuracy than standard radiation therapy. The intent of this review is to describe the development and evaluate the possible benefit of AI tools integration into the radiation oncology workflow for SBRT automation. The selected papers were subdivided into four sections, representative of the whole radiotherapy process: 'AI in SBRT target and organs at risk contouring', 'AI in SBRT planning', 'AI during the SBRT delivery', and 'AI for outcome prediction after SBRT'. Each section summarises the challenges, as well as limits and needs for improvement to achieve better integration of AI tools in the clinical workflow.

Novel Harmonization Method for Multi-Centric Radiomic Studies in Non-Small Cell Lung Cancer.

The purpose of this multi-centric work was to investigate the relationship between radiomic features extracted from pre-treatment computed tomography (CT), positron emission tomography (PET) imaging, and clinical outcomes for stereotactic body radiation therapy (SBRT) in early-stage non-small cell lung cancer (NSCLC). One-hundred and seventeen patients who received SBRT for early-stage NSCLC were retrospectively identified from seven Italian centers. The tumor was identified on pre-treatment free-breathing CT and PET images, from which we extracted 3004 quantitative radiomic features. The primary outcome was 24-month progression-free-survival (PFS) based on cancer recurrence (local/non-local) following SBRT. A harmonization technique was proposed for CT features considering lesion and contralateral healthy lung tissues using the LASSO algorithm as a feature selector. Models with harmonized CT features (B models) demonstrated better performances compared to the ones using only original CT features (C models). A linear support vector machine (SVM) with harmonized CT and PET features (A1 model) showed an area under the curve (AUC) of 0.77 (0.63-0.85) for predicting the primary outcome in an external validation cohort. The addition of clinical features did not enhance the model performance. This study provided the basis for validating our novel CT data harmonization strategy, involving delta radiomics. The harmonized radiomic models demonstrated the capability to properly predict patient prognosis.

Deep learning based time-to-event analysis with PET, CT and joint PET/CT for head and neck cancer prognosis.

OBJECTIVES: Recent studies have shown that deep learning based on pre-treatment positron emission tomography (PET) or computed tomography (CT) is promising for distant metastasis (DM) and overall survival (OS) prognosis in head and neck cancer (HNC). However, lesion segmentation is typically required, resulting in a predictive power susceptible to variations in primary and lymph node gross tumor volume (GTV) segmentation. This study aimed at achieving prognosis without GTV segmentation, and extending single modality prognosis to joint PET/CT to allow investigating the predictive performance of combined- compared to single-modality inputs. METHODS: We employed a 3D-Resnet combined with a time-to-event outcome model to incorporate censoring information. We focused on the prognosis of DM and OS for HNC patients. For each clinical endpoint, five models with PET and/or CT images as input were compared: PET-GTV, PET-only, CT-GTV, CT-only, and PET/CT-GTV models, where -GTV indicates that the corresponding images were masked using the GTV contour. Publicly available delineated CT and PET scans from 4 different Canadian hospitals (293) and the MAASTRO clinic (74) were used for training by 3-fold cross-validation (CV). For independent testing, we used 110 patients from a collaborating institution. The predictive performance was evaluated via Harrell's Concordance Index (HCI) and Kaplan-Meier curves. RESULTS: In a 5-year time-to-event analysis, all models could produce CV HCIs with median values around 0.8 for DM and 0.7 for OS. The best performance was obtained with the PET-only model, achieving a median testing HCI of 0.82 for DM and 0.69 for OS. Compared with the PET/CT-GTV model, the PET-only still had advantages of up to 0.07 in terms of testing HCI. The Kaplan-Meier curves and corresponding log-rank test results also demonstrated significant stratification capability of our models for the testing cohort. CONCLUSION: Deep learning-based DM and OS time-to-event models showed predictive capability and could provide indications for personalized RT. The best predictive performance achieved by the PET-only model suggested GTV segmentation might be less relevant for PET-based prognosis.

Breast Hypoplasia and Decreased Lactation From Radiation Therapy in Survivors of Pediatric Malignancy: A PENTEC Comprehensive Review.

PURPOSE: Breast hypoplasia and impaired lactation are poorly studied sequelae of chest radiation therapy (RT) in children. The Pediatric Normal Tissue Effects in the Clinic female breast task force aimed to quantitate the radiation dose-volume effects on these endpoints. METHODS AND MATERIALS: A literature search was conducted of peer-reviewed manuscripts evaluating breast hypoplasia and lactation after chest RT in children, yielding 789 abstracts. Only 2 studies on children irradiated at <4 years of age for angioma of the breast provided dosimetric data correlated with breast hypoplasia. For patients who received brachytherapy, the dose was converted to external beam RT in equivalent 2 Gy fractions (DEBRT), although the limitations of this type of mathematical conversion need to be recognized. We calculated relative risks (RR) and 95% confidence intervals (95% CIs) based on these data. Only 1 study was relevant to the lactation endpoint, in which patients were given RT for Hodgkin lymphoma at age 14 to 40 years. RESULTS: The 3 studies involved 206 patients in total. In patients <4 years old at the time of RT, the prevalence of patient-perceived breast hypoplasia was 38% (RR 2.5; 95% CI, 1.3-4.6) after DEBRT of <0.34 Gy, 61% (RR 4.0; 95% CI, 2.1-7.4) after DEBRT 0.34-0.97 Gy, and 97% (RR 6.3; 95% CI, 3.6-10.8) after DEBRT ≥0.97 Gy to the breast anlage. A simple linear regression model (r = 0.72; P < .001) showed that the treated breast was smaller than the untreated breast by 13% at DEBRT = 0.5 Gy, 20% at DEBRT = 1 Gy, 32% at DEBRT = 2 Gy, 51% at DEBRT = 4 Gy, 66% at DEBRT = 6 Gy, 79% at DEBRT = 8 Gy, and 90% at DEBRT = 10 Gy. The risk of unsuccessful breastfeeding was 39% after a median mediastinal dose of 41 Gy, compared with 21% in a sibling control group (P = .04). RT dose of ≥42 Gy was not associated with less breastfeeding success compared with <42 Gy, and data on lower doses were unavailable. CONCLUSIONS: Based on extremely limited data, young adults exposed to thoracic RT as children seem to be at significant risk of breast hypoplasia and impaired lactation. Doses as low as 0.3 Gy to immature breasts can cause breast hypoplasia. Additional studies are needed to quantify dose and technique effects with modern RT indications. Prospective collection of clinical outcomes and dosimetric factors would enhance our understanding of RT-induced breast hypoplasia and impaired lactation.

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 (RadCT ), radiomics CT and BED (RadCT,BED ), deep learning (DL) CT (DLCT ), and DL CT and BED (DLCT,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 RadCT,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 DLCT,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.

A Multicentre Evaluation of Dosiomics Features Reproducibility, Stability and Sensitivity.

Dosiomics is a texture analysis method to produce dose features that encode the spatial 3D distribution of radiotherapy dose. Dosiomic studies, in a multicentre setting, require assessing the features' stability to dose calculation settings and the features' capability in distinguishing different dose distributions. Dose distributions were generated by eight Italian centres on a shared image dataset acquired on a dedicated phantom. Treatment planning protocols, in terms of planning target volume coverage and dose-volume constraints to the organs at risk, were shared among the centres to produce comparable dose distributions for measuring reproducibility/stability and sensitivity of dosiomic features. In addition, coefficient of variation (CV) was employed to evaluate the dosiomic features' variation. We extracted 38,160 features from 30 different dose distributions from six regions of interest, grouped by four features' families. A selected group of features (CV < 3 for the reproducibility/stability studies, CV > 1 for the sensitivity studies) were identified to support future multicentre studies, assuring both stable features when dose distributions variation is minimal and sensitive features when dose distribution variations need to be clearly identified. Dosiomic is a promising tool that could support multicentre studies, especially for predictive models, and encode the spatial and statistical characteristics of the 3D dose distribution.

miR-9 modulates and predicts the response to radiotherapy and EGFR inhibition in HNSCC.

Radiotherapy (RT) plus the anti-EGFR monoclonal antibody Cetuximab (CTX) is an effective combination therapy for a subset of head and neck squamous cell carcinoma (HNSCC) patients. However, predictive markers of efficacy are missing, resulting in many patients treated with disappointing results and unnecessary toxicities. Here, we report that activation of EGFR upregulates miR-9 expression, which sustains the aggressiveness of HNSCC cells and protects from RT-induced cell death. Mechanistically, by targeting KLF5, miR-9 regulates the expression of the transcription factor Sp1 that, in turn, stimulates tumor growth and confers resistance to RT+CTX in vitro and in vivo. Intriguingly, high miR-9 levels have no effect on the sensitivity of HNSCC cells to cisplatin. In primary HNSCC, miR-9 expression correlated with Sp1 mRNA levels and high miR-9 expression predicted poor prognosis in patients treated with RT+CTX. Overall, we have discovered a new signaling axis linking EGFR activation to Sp1 expression that dictates the response to combination treatments in HNSCC. We propose that miR-9 may represent a valuable biomarker to select which HNSCC patients might benefit from RT+CTX therapy.

Artificial intelligence applications in medical imaging: A review of the medical physics research in Italy.

PURPOSE: To perform a systematic review on the research on the application of artificial intelligence (AI) to imaging published in Italy and identify its fields of application, methods and results. MATERIALS AND METHODS: A Pubmed search was conducted using terms Artificial Intelligence, Machine Learning, Deep learning, imaging, and Italy as affiliation, excluding reviews and papers outside time interval 2015-2020. In a second phase, participants of the working group AI4MP on Artificial Intelligence of the Italian Association of Physics in Medicine (AIFM) searched for papers on AI in imaging. RESULTS: The Pubmed search produced 794 results. 168 studies were selected, of which 122 were from Pubmed search and 46 from the working group. The most used imaging modality was MRI (44%) followed by CT(12%) ad radiography/mammography (11%). The most common clinical indication were neurological diseases (29%) and diagnosis of cancer (25%). Classification was the most common task for AI (57%) followed by segmentation (16%). 65% of studies used machine learning and 35% used deep learning. We observed a rapid increase of research in Italy on artificial intelligence in the last 5 years, peaking at 155% from 2018 to 2019. CONCLUSIONS: We are witnessing an unprecedented interest in AI applied to imaging in Italy, in a diversity of fields and imaging techniques. Further initiatives are needed to build common frameworks and databases, collaborations among different types of institutions, and guidelines for research on AI.

Neurocognitive Effects and Necrosis in Childhood Cancer Survivors Treated With Radiation Therapy: A PENTEC Comprehensive Review.

PURPOSE: A PENTEC review of childhood cancer survivors who received brain radiation therapy (RT) was performed to develop models that aid in developing dose constraints for RT-associated central nervous system (CNS) morbidities. METHODS AND MATERIALS: A comprehensive literature search, through the PENTEC initiative, was performed to identify published data pertaining to 6 specific CNS toxicities in children treated with brain RT. Treatment and outcome data on survivors were extracted and used to generate normal tissue complication probability (NTCP) models. RESULTS: The search identified investigations pertaining to 2 of the 6 predefined CNS outcomes: neurocognition and brain necrosis. For neurocognition, models for 2 post-RT outcomes were developed to (1) calculate the risk for a below-average intelligence quotient (IQ) (IQ <85) and (2) estimate the expected IQ value. The models suggest that there is a 5% risk of a subsequent IQ <85 when 10%, 20%, 50%, or 100% of the brain is irradiated to 35.7, 29.1, 22.2, or 18.1 Gy, respectively (all at 2 Gy/fraction and without methotrexate). Methotrexate (MTX) increased the risk for an IQ <85 similar to a generalized uniform brain dose of 5.9 Gy. The model for predicting expected IQ also includes the effect of dose, age, and MTX. Each of these factors has an independent, but probably cumulative effect on IQ. The necrosis model estimates a 5% risk of necrosis for children after 58.9 Gy or 59.9 Gy (2 Gy/fraction) to any part of the brain if delivered as primary RT or reirradiation, respectively. CONCLUSIONS: This PENTEC comprehensive review establishes objective relationships between patient age, RT dose, RT volume, and MTX to subsequent risks of neurocognitive injury and necrosis. A lack of consistent RT data and outcome reporting in the published literature hindered investigation of the other predefined CNS morbidity endpoints.

Distant metastasis time to event analysis with CNNs in independent head and neck cancer cohorts.

Deep learning models based on medical images play an increasingly important role for cancer outcome prediction. The standard approach involves usage of convolutional neural networks (CNNs) to automatically extract relevant features from the patient's image and perform a binary classification of the occurrence of a given clinical endpoint. In this work, a 2D-CNN and a 3D-CNN for the binary classification of distant metastasis (DM) occurrence in head and neck cancer patients were extended to perform time-to-event analysis. The newly built CNNs incorporate censoring information and output DM-free probability curves as a function of time for every patient. In total, 1037 patients were used to build and assess the performance of the time-to-event model. Training and validation was based on 294 patients also used in a previous benchmark classification study while for testing 743 patients from three independent cohorts were used. The best network could reproduce the good results from 3-fold cross validation [Harrell's concordance indices (HCIs) of 0.78, 0.74 and 0.80] in two out of three testing cohorts (HCIs of 0.88, 0.67 and 0.77). Additionally, the capability of the models for patient stratification into high and low-risk groups was investigated, the CNNs being able to significantly stratify all three testing cohorts. Results suggest that image-based deep learning models show good reliability for DM time-to-event analysis and could be used for treatment personalisation.

Partial prostate re-irradiation for the treatment of isolated local recurrence of prostate cancer in patients previously treated with primary external beam radiotherapy: short-term results of a monocentric study.

Many different therapeutic options are available for locally recurrent prostate cancer (PCa). However, standard treatment has not yet been established. We conducted a partial prostate re-irradiation (PPR) program for the treatment of isolated and limited-size intraprostatic recurrences, in patients who previously underwent external beam radiation therapy (EBRT) as primary treatment for prostatic cancer (PCa). The analysis of this experience in terms of feasibility, toxicity, and efficacy is reported. The inclusion criteria of this retrospective analysis were: previous definitive EBRT, evidence of biochemical recurrence, radiological detection of isolated local relapse, and PPR as local salvage therapy. Gastrointestinal (GI) and genitourinary (GU) toxicities were registered according to the RTOG/EORTC criteria. Between July 2012 and May 2019, 44 patients were treated with PPR. All patients completed the planned treatment. The median follow-up was 25.4 months. Tumor progression was observed in 18 patients (40.9%). Two-year local control, biochemical failure-, and clinical relapse-free survival rates were 90.1%, 58.3%, and 67.9%, respectively. The occurrence of biochemical failure after PPR is lower for patients with the time interval between the primary EBRT and first biochemical failure >4 years; local control results strongly associated with a biologically effective dose (BED) at first EBRT >177 Gy. No acute grade 3 or greater toxic events were observed. Two late grade 3 GU toxicities were reported. Although retrospective in design, our study indicates that PPR appears as a feasible, well-tolerated, and effective salvage treatment for isolated local PCa recurrence. Long term data are required in order to confirm these results.

Radiomics and deep learning in lung cancer.

Lung malignancies have been extensively characterized through radiomics and deep learning. By providing a three-dimensional characterization of the lesion, models based on radiomic features from computed tomography (CT) and positron-emission tomography (PET) have been developed to detect nodules, distinguish malignant from benign lesions, characterize their histology, stage, and genotype. Deep learning models have been applied to automatically segment organs at risk in lung cancer radiotherapy, stratify patients according to the risk for local and distant recurrence, and identify patients candidate for molecular targeted therapy and immunotherapy. Moreover, radiomics has also been applied successfully to predict side effects such as radiation- and immunotherapy-induced pneumonitis and differentiate lung injury from recurrence. Radiomics could also untap the potential for further use of the cone beam CT acquired for treatment image guidance, four-dimensional CT, and dose-volume data from radiotherapy treatment plans. Radiomics is expected to increasingly affect the clinical practice of treatment of lung tumors, optimizing the end-to-end diagnosis-treatment-follow-up chain. The main goal of this article is to provide an update on the current status of lung cancer radiomics.