CT respiratory motion synthesis using joint supervised and adversarial learning

Objective.Four-dimensional computed tomography (4DCT) imaging consists in reconstructing a CT acquisition into multiple phases to track internal organ and tumor motion. It is commonly used in radiotherapy treatment planning to establish planning target volumes. However, 4DCT increases protocol complexity, may not align with patient breathing during treatment, and lead to higher radiation delivery.Approach.In this study, we propose a deep synthesis method to generate pseudo respiratory CT phases from static images for motion-aware treatment planning. The model produces patient-specific deformation vector fields (DVFs) by conditioning synthesis on external patient surface-based estimation, mimicking respiratory monitoring devices. A key methodological contribution is to encourage DVF realism through supervised DVF training while using an adversarial term jointly not only on the warped image but also on the magnitude of the DVF itself. This way, we avoid excessive smoothness typically obtained through deep unsupervised learning, and encourage correlations with the respiratory amplitude.Main results.Performance is evaluated using real 4DCT acquisitions with smaller tumor volumes than previously reported. Results demonstrate for the first time that the generated pseudo-respiratory CT phases can capture organ and tumor motion with similar accuracy to repeated 4DCT scans of the same patient. Mean inter-scans tumor center-of-mass distances and Dice similarity coefficients were 1.97 mm and 0.63, respectively, for real 4DCT phases and 2.35 mm and 0.71 for synthetic phases, and compares favorably to a state-of-the-art technique (RMSim).Significance.This study presents a deep image synthesis method that addresses the limitations of conventional 4DCT by generating pseudo-respiratory CT phases from static images. Although further studies are needed to assess the dosimetric impact of the proposed method, this approach has the potential to reduce radiation exposure in radiotherapy treatment planning while maintaining accurate motion representation. Our training and testing code can be found athttps://github.com/cyiheng/Dynagan.

PET-Based Radiogenomics Supports KEAP1/NFE2L2 Pathway Targeting for Non-Small Cell Lung Cancer Treated with Curative Radiotherapy

PURPOSE/OBJECTIVE(S)

Radiotherapy (RT) is one of the major treatment options for localized lung cancer. Either delivered in normo- or moderate/highly hypofractionated regimens, use of RT is increasing especially thanks to the development of stereotactic radiotherapy (SBRT). RT is associated with a higher risk of local relapse when compared with surgery but with a preferable toxicity profile. The KEAP1/NFE2L2 mutational status has been proven as significantly correlated with local relapse in patients treated with RT. Several (33) transcriptomic features were previously identified as dependent from the KEAP1/NFE2L2 mutational status. However, these genetic and transcriptomic tests are rarely performed because of their cost and lack of availability. Prediction of the KEAP1/NFE2L2 mutational status on non-invasive modalities such as imaging could help in further personalizing each therapeutic strategy.

MATERIALS/METHODS

Due to the small size of patients with both mutation status (MutKEAP1-NFE2L2) and PET/CT, a first model (RNASeq) predicting the mutation status (MutRNASeq) using the 33 previously identified transcriptomic features was developed on patients from the TCGA-LUSC, TCGA-LUAD, CPTAC-LSCC and CPTAC-LUAD cohorts (770 patients) and externally validated on the NSCLC-Radiogenomics cohort (117 patients). Narrowing the patients to those with an available PET/CT, a second model (RNAPET) was then built and internally validated to predict the previously MutRNASeq probability using PET/CT-extracted radiomics features. The RNAPET model was then validated on an external cohort of 151 patients treated with curative radiotherapy for a localized non-small cell lung cancer (VMAT cohort). For each model, features were combined using a neural network approach (Multilayer Perceptron) within a statistical software modeler. Performances were evaluated based on the ROC-features as well as decision curve analysis.

RESULTS

The RNASeq model resulted in a C-Index of 0.82, Sensitivity (Se) of 70.3% and Specificity (Sp) of 93.4% in the validation cohort. Regarding the PET/CT-based prediction on a training cohort of 101 patients, the retained RNAPET model resulted in an AUC of 0.90 (p < 0.001). With a probability threshold of 20% and applied to the testing cohort, the RNAPET model achieved a C-Index of 0.7 with respective Se/Sp of 60.0% and 80.9% for the prediction of the MutRNASeq. The same radiomics model was validated on the VMAT cohort as patients were significantly stratified based on their risk of locoregional (LR) relapse with a hazard ratio of 2.61 (p = 0.02).

CONCLUSION

Our three-step approach enables the prediction of the MutKEAP1-NFE2L2 using PET/CT-extracted radiomics features and efficiently classified patient at risk of LR relapse in an external cohort treated with radiotherapy. This first evidence should be further evaluated on larger cohorts, and implemented in LR risk prediction models.

Radiotherapy modification based on artificial intelligence and radiomics applied to (18F)-fluorodeoxyglucose positron emission tomography/computed tomography

Over the last decades, the refinement of radiation therapy techniques has been associated with an increasing interest for individualized radiation therapy with the aim of increasing or maintaining tumor control and reducing radiation toxicity. Developments in artificial intelligence (AI), particularly machine learning and deep learning, in imaging sciences, including nuclear medecine, have led to significant enthusiasm for the concept of "rapid learning health system". AI combined with radiomics applied to (18F)-fluorodeoxyglucose positron emission tomography/computed tomography ([18F]-FDG PET/CT) offers a unique opportunity for the development of predictive models that can help stratify each patient's risk and guide treatment decisions for optimal outcomes and quality of life of patients treated with radiation therapy. Here we present an overview of the current contribution of AI and radiomics-based machine learning models applied to (18F)-FDG PET/CT in the management of cancer treated by radiation therapy.

Impact of age on the Quadrella index assessing oncological and functional results after prostate brachytherapy: A 6-year analysis

Purpose

As the oncological results of prostate brachytherapy (BT) are excellent for low-risk (LR) or favorable intermediate-risk (FIR) prostate cancer (PCa), evaluating the side effects has become a major issue, especially for young men. The objective of the study was to compare the oncologic and functional results of BT using Quadrella index for patients aged 60 or less compared with older patients.

Material and methods

From June, 2007 to June, 2017, 222 patients, including 70 ≤ 60 years old and 152 > 60 years old, underwent BT for LR-FIR PCa, with good erectile function at baseline according to International Index of Erectile Function-5 (IIEF-5) > 16. Quadrella index was achieved under the following circumstances: 1) Absence of biological recurrence (Phoenix criteria); 2) Absence of erectile dysfunction (ED) (IIEF-5 > 16); 3) No urinary toxicity (international prostate score symptom) IPSS < 15 or IPSS > 15, and ΔIPSS < 5; 4) No rectal toxicity (RT) (Radiation Therapy Oncology Group, RTOG = 0). Patients were treated on demand with phosphodiesterase inhibitors (PDE5i) post-operatively.

Results

The Quadrella index was satisfied for about 40-80% of patients ≤ 60 years vs. 33-46% for older patients during 6-year follow-up (significant difference from the second year). At year 5, 100% of evaluable patients aged ≤ 60 and 91.8% > 60 (p = 0.29) reached Phoenix criteria. The criterion of ED (IIEF-5 < 16) largely explained the validity rate of Quadrella alone. There was no ED for 67.2-81.4% of patients ≤ 60 years compared with 40.0-56.1% for patients > 60 (significant difference since year 4 in favor of young men). After two years of follow-up, more than 90% of patients in both the groups showed neither urinary nor rectal toxicities.

Conclusions

For young men displaying LR-FIR PCa, BT appears to be a first-class therapeutic option, as the oncological results were at least equivalent to those of older patients with good long-term tolerance.

Recommendations for stereotactic body radiation therapy for spine and non-spine bone metastases. A GETUG (French society of urological radiation oncolgists) consensus using a national two-round modified Delphi survey

Background and purpose

The relevance of metastasis-directed stereotactic body radiation therapy (SBRT) remains to be demonstrated through phase III trials. Multiple SBRT procedures have been published potentially resulting in a disparity of practices. Therefore, the french society of urological radiation oncolgists (GETUG) recognized the need for joint expert consensus guidelines for metastasis-directed SBRT in order to standardize practice in trials carried out by the group.

Materials and methods

After a comprehensive literature review, 97 recommendation statements were created regarding planning and delivery of spine bone (SBM) and non-spine bone metastases (NSBM) SBRT. These statements were then submitted to a national online two-round modified Delphi survey among main GETUG investigators. Consensus was achieved if a statement received ≥ 75 % agreements, a trend to consensus being defined as 65-74 % agreements. Any statement without consensus at round one was re-submitted in round two.

Results

Twenty-one out of 29 (72.4%) surveyed experts responded to both rounds. Seventy-five statements achieved consensus at round one leaving 22 statements needing a revote of which 16 achieved consensus and 5 a trend to consensus. The final rate of consensus was 91/97 (93.8%). Statements with no consensus concerned patient selection (3/19), dose and fractionation (1/11), prescription and dose objectives (1/9) and organs at risk delineation (1/15). The voting resulted in the writing of step-by-step consensus guidelines.

Conclusion

Consensus guidelines for SBM and NSBM SBRT were agreed upon using a validated modified Delphi approach. These guidelines will be used as per-protocole recommendations in ongoing and further GETUG clinical trials.

Impact of radiotherapy on survival in resected or unresectable anaplastic thyroid carcinomas, a Rare Cancer Network study

PURPOSE

Anaplastic thyroid carcinomas (ATC) are a heterogenous group of tumors of overall dismal prognosis. We designed models to identify relevant prognostic factors of survival of irradiated ATC patients including radiotherapy modalities (field size, dose).

MATERIAL AND METHODS

Between 2000 and 2017, 166 ATC patients' treatments were divided into surgery and postoperative radiotherapy (poRT) or definitive radiotherapy (RT). Multiple imputation approach was used for missing data. Prognostic factors were identified using Lasso-penalized Cox modelling and predicted risk scores were built.

RESULTS

Patients undergoing RT (n=70) had more adverse patient and disease characteristics than those undergoing poRT (n=96). Corresponding median survival rates were 5.4 and 12.1 months, respectively. PoRT patients undergoing poRT more likely received extended-field radiotherapy with prophylactic nodal irradiation, but rather received platinum- vs. adriamycin-based chemoradiotherapy. Radiotherapy was conventionally fractionated, delivered >60Gy in 51.9% and 61.7% and used extended fields in 88.5% and 71.2% of patients with poRT or RT. Radiotherapy interruption rates for toxicity were similar in the two groups. The best poRT-group model identified age>45yo, PS≥1, pathologic tumor stage≥pT4b,>N1 and R2 resection as poor prognostic factors. The best RT-group model (C-index of 0.72) identified PS≥3,>N1 and extended-field radiotherapy with prophylactic nodal irradiation (as opposed to tumour-bed irradiation only) as poor prognostic factors.

CONCLUSION

In patients undergoing poRT, radiotherapy parameters had little influence over their survival irrespective of patient, disease characteristics, and quality of resection. In patients undergoing RT, extended-field radiotherapy improved survival in addition to PS and nodal stage.

Predicting response to radiotherapy of head and neck squamous cell carcinoma using radiomics from cone-beam CT images

INTRODUCTION

Radiotherapy (RT) for head and neck cancer is now guided by cone-beam computed tomography (CBCT). We aim to identify a CBCT radiomic signature predictive of progression to RT.

MATERIAL AND METHODS

A cohort of 93 patients was split into training (n = 60) and testing (n = 33) sets. A total of 88 features were extracted from the gross tumor volume (GTV) on each CBCT. Receiver operating characteristic (ROC) curves were used to determine the power of each feature at each week of treatment to predict progression to radio(chemo)therapy. Only features with AUC > 0.65 at each week were pre-selected. Absolute differences were calculated between features from each weekly CBCT and baseline CBCT1 images. The smallest detectable change (C = 1.96 × SD, SD being the standard deviation of differences between feature values calculated on CBCT1 and CBCTn) with its confidence interval (95% confidence interval [CI]) was determined for each feature. The features for which the change was larger than C for at least 5% of patients were then selected. A radiomics-based model was built at the time-point that showed the highest AUC and compared with models relying on clinical variables.

RESULTS

Seven features had an AUC > 0.65 at each week, and six exhibited a change larger than the predefined CI 95%. After exclusion of inter-correlated features, only one parameter remains, Coarseness. Among clinical variable, only hemoglobin value was significant. AUC for predicting the treatment response were 0.78 (p = .006), 0.85 (p < .001), and 0.99 (p < .001) for clinical, CBCT4-radiomics (Coarseness) and clinical + radiomics based models respectively. The mean AUC of this last model on a 5-fold cross-validation was 0.80 (±0.09). On the testing cohort, the best prediction was given by the combined model (balanced accuracy [BAcc] 0.67 , p < .001).

CONCLUSIONS

We described a feature selection methodology for delta-radiomics that is able to select reproducible features which are informative due to their change during treatment. A selected delta radiomics feature may improve clinical-based prediction models.

Radiotherapy and pregnancy

We present the updated recommendations of the French society for radiation oncology on radiotherapy and pregnancy. The occurrence of cancer during pregnancy is a rare event (approximately 1 in 1000 pregnancies). The risks for the embryo or the foetus depend on the gestational age at the time of irradiation. The main risks are malformations with microcephaly and mental retardation. There is also a risk of radiation-induced cancer in the unborn child. In the case of only supradiaphragmatic irradiation, radiotherapy can be performed most often in pregnant women without risk to the foetus. On the other hand, in the case of an indication for subdiaphragmatic irradiation, therapeutic termination of the pregnancy should be proposed. In all cases, when radiotherapy is chosen, a phantom estimation of the dose delivered to the foetus, confirmed by in vivo measurement, is recommended. Conformational radiotherapy is the preferred technique because of the lower dose delivered to the foetus (except in tumour locations where other techniques such as IMRT are recommended).

[Salvage radiotherapy after prostatectomy - what do do in case of elevated post operative PSA or macroscopic relapse?]

The management of the postoperative biological relapse of prostate cancer is most often based on salvage radiotherapy (RT) and a variable duration of hormone therapy (HT) in addition. The indications for RT±HT become more consensual for the adjuvant postoperative situation or in the event of a rising PSA level after a period where an undetectable PSA level was achieved. On the other hand, in the event of detectable PSA immediately postoperatively or in the event of a biological recurrence with macroscopic relapse in the prostate bed seen on conventional imaging assessment, the treatment options are still being evaluated. This article will describe these 2 situations and their current management but also will come through assessments with the contribution of modern imaging and new treatment options in terms of RT dose and RT±HT combination.

Postoperative radiotherapy in prostate cancer: Dose and volumes

Approximately thirty percent of patients experience biochemical recurrence after radical prostatectomy for prostate cancer. Early salvage radiotherapy has recently become a standard of care in this setting. The purpose of this review is first to summarize current knowledge in terms of dose to the prostate bed in light of the recent SAKK 09/10 randomized phase III trial results. The evidence on moderate hypofractionation will also be discussed whereas extreme hypofractionation remains highly investigational. Regarding target volumes, several different guidelines have been published to address the need for standardization of postoperative target delineation. The recent GFRU (Groupe Francophone de Radiothérapie Urologique) recommendations could represent an international consensus.

Histosurgical mapping of endoscopic endonasal surgery of sinonasal tumours to improve radiotherapy guidance

PURPOSE

Endoscopic endonasal surgery (EES) is becoming a standard for most malignant sinonasal tumours. Margin analysis after piecemeal resection is complex and optimally relies on accurate histosurgical mapping. Postoperative radiotherapy may be adapted based on margin assessment mapping to reduce the dose to some sinonasal subvolumes. We assessed the use of histosurgical mapping by radiation oncologists (RO).

MATERIAL AND METHODS

A French practice survey was performed across 29 ENT expert RO (2 did not answer) regarding integration of information on EES, as well as quality of operative and pathology reportsto refine radiotherapy planning after EES. This was assessed through an electronic questionnaire.

RESULTS

EES was ubiquitously performed in France. Operative and pathology reports yielded accurate description of EES samples according to 66.7% of interviewed RO. Accuracy of margin assessment was however insufficient according to more than 40.0% of RO. Additional margins/biopsies of the operative bed were available in 55.2% (16/29) of the centres. In the absence of additional margins, quality of resection after EES was considered as microscopically incomplete in 48.3% or dubious in 48.3% of RO. As performed, histosurgical mapping allowed radiotherapy dose and volumes adaptation according to 26.3% of RO only.

CONCLUSIONS

Standardized histosurgical mapping with margin and additional margin analysis could be more systematic. Advantages of accurate EES reporting could be dose painting radiotherapy to further decrease morbidity in sinonasal tumours.

Integration of functional imaging in brachytherapy

Functional imaging allows the evaluation of numerous biological properties that could be considered at all steps of the therapeutic management of patients treated with brachytherapy. Indeed, it enables better initial staging of the disease, and some parameters may also be used as predictive biomarkers for treatment response, allowing better selection of patients eligible for brachytherapy. It may also improve the definition of target volumes with the aim of dose escalations by dose-painting. Finally, it could be useful during the follow-up to assess response to treatment. In this review, we report how functional imaging is integrated at the present time during the brachytherapy procedure, and what are its potential future contributions in the main tumour locations where brachytherapy is recommended. Functional imaging has great potential in the contact of brachytherapy, but still, several issues remain to be resolved before integrating it into clinical practice, especially as a biomarker or in dose painting strategies.

Radiomics analysis of 3D dose distributions to predict toxicity of radiotherapy for lung cancer

PURPOSE

(Chemo)-radiotherapy (RT) is the gold standard treatment for patients with locally advanced lung cancer non accessible for surgery. However, current toxicity prediction models rely on clinical and dose volume histograms (DVHs) and remain unsufficient. The goal of this work is to investigate the added predictive value of the radiomics approach applied to dose maps regarding acute and late toxicities in both the lungs and esophagus.

METHODS

Acute and late toxicities scored using the CTCAE v4.0 were retrospectively collected on patients treated with RT in our institution. Radiomic features were extracted from 3D dose maps considering Gy values as grey-levels in images. DVH and usual clinical factors were also considered. Three toxicity prediction models (clinical only, clinical + DVH and combined, i.e., including clinical + DVH + radiomics) were incrementally trained using a neural network on 70% of the patients for prediction of grade ≥2 acute and late pulmonary toxicities (APT/LPT) and grade ≥2 acute esophageal toxicities (AET). After bootstrapping (n = 1000), optimal cut-off values were determined based on the Youden Index. The trained models were then evaluated in the remaining 30% of patients using balanced accuracy (BAcc).

RESULTS

167 patients were treated from 2015 to 2018: 78% non small-cell lung cancers, 14% small-cell lung cancers and 8% other histology with a median age at treatment of 66 years. Respectively, 22.2%, 16.8% and 30.0% experienced APT, LPT and AET. In the training set (n = 117), the corresponding BAcc for clinical only/clinical + DVH/combined were 0.68/0.79/0.92, 0.66/0.77/0.87 and 0.68/0.73/0.84. In the testing evaluation (n = 50), these trained models obtained a corresponding BAcc of 0.69/0.69/0.92, 0.76/0.80/0.89 and 0.58/0.73/0.72.

CONCLUSION

In patients with a lung cancer treated with RT, radiomic features extracted from 3D dose maps seem to surpass usual models based on clinical factors and DVHs for the prediction of APT and LPT.

Performance comparison of modified ComBat for harmonization of radiomic features for multicenter studies

Multicenter studies are needed to demonstrate the clinical potential value of radiomics as a prognostic tool. However, variability in scanner models, acquisition protocols and reconstruction settings are unavoidable and radiomic features are notoriously sensitive to these factors, which hinders pooling them in a statistical analysis. A statistical harmonization method called ComBat was developed to deal with the "batch effect" in gene expression microarray data and was used in radiomics studies to deal with the "center-effect". Our goal was to evaluate modifications in ComBat allowing for more flexibility in choosing a reference and improving robustness of the estimation. Two modified ComBat versions were evaluated: M-ComBat allows to transform all features distributions to a chosen reference, instead of the overall mean, providing more flexibility. B-ComBat adds bootstrap and Monte Carlo for improved robustness in the estimation. BM-ComBat combines both modifications. The four versions were compared regarding their ability to harmonize features in a multicenter context in two different clinical datasets. The first contains 119 locally advanced cervical cancer patients from 3 centers, with magnetic resonance imaging and positron emission tomography imaging. In that case ComBat was applied with 3 labels corresponding to each center. The second one contains 98 locally advanced laryngeal cancer patients from 5 centers with contrast-enhanced computed tomography. In that specific case, because imaging settings were highly heterogeneous even within each of the five centers, unsupervised clustering was used to determine two labels for applying ComBat. The impact of each harmonization was evaluated through three different machine learning pipelines for the modelling step in predicting the clinical outcomes, across two performance metrics (balanced accuracy and Matthews correlation coefficient). Before harmonization, almost all radiomic features had significantly different distributions between labels. These differences were successfully removed with all ComBat versions. The predictive ability of the radiomic models was always improved with harmonization and the improved ComBat provided the best results. This was observed consistently in both datasets, through all machine learning pipelines and performance metrics. The proposed modifications allow for more flexibility and robustness in the estimation. They also slightly but consistently improve the predictive power of resulting radiomic models.

Kilovoltage intrafraction monitoring during normofractionated prostate cancer radiotherapy

PURPOSE

During radiotherapy (RT) for prostate cancer (PCa), interfraction and intrafraction movements can lead to decreased target dose coverage and unnecessary over-exposure of organs at risk. New image-guided RT techniques accuracy allows planning target volume (PTV) margins reduction. We aim to assess the feasibility of a kilovoltage intrafraction monitoring (KIM) to track the prostate during RT.

METHODS AND MATERIALS

Between November 2017 and April 2018, 44 consecutive patients with PCa were included in an intrafraction prostate motion study using the Truebeam Auto Beam Hold® tracking system (Varian Medical Systems, United State) triggered by gold fiducials localization on kilovoltage (kV) imaging. A 5-mm PTV was considered. A significant gating event (SGE) was defined as the occurrence of an automatic beam interruption requiring patient repositioning following the detection of one fiducial outside a 5-mm target area around the marker during more than 45seconds.

RESULTS

Six patients could not benefit from the KIM because of technical issues (loss of one fiducial marker=1, hip prosthesis=4, morbid obesity causing table movements=1). The mean rate of SGE per patient was 14±19%, and the fraction average delivery time was increased by 146±86seconds. For a plan of 39 fractions of 2Gy, the additional radiation dose increased by 0.13±0.09Gy. The mean rates of SGE were 2% and 18% (P=0.002) in patients with planned fraction<90 and>90seconds respectively, showing that duration of the session strongly interfered with prostate intrafraction movements. No other significant clinical and technical parameter was correlated with the occurrence of SGE.

CONCLUSION

Automated intrafraction kV imaging can effectively perform autobeam holds due to intrafraction movement of the prostate in the large majority of patients. The additional radiation dose and delivery time are acceptable. This technique may be a cost-effective alternative to electromagnetic transponder guidance.