Deep Hybrid Learning Prediction of Patient-Specific Quality Assurance in Radiotherapy: Implementation in Clinical Routine.

BACKGROUND: Arc therapy allows for better dose deposition conformation, but the radiotherapy plans (RT plans) are more complex, requiring patient-specific pre-treatment quality assurance (QA). In turn, pre-treatment QA adds to the workload. The objective of this study was to develop a predictive model of Delta4-QA results based on RT-plan complexity indices to reduce QA workload. METHODS: Six complexity indices were extracted from 1632 RT VMAT plans. A machine learning (ML) model was developed for classification purpose (two classes: compliance with the QA plan or not). For more complex locations (breast, pelvis and head and neck), innovative deep hybrid learning (DHL) was trained to achieve better performance. RESULTS: For not complex RT plans (with brain and thorax tumor locations), the ML model achieved 100% specificity and 98.9% sensitivity. However, for more complex RT plans, specificity falls to 87%. For these complex RT plans, an innovative QA classification method using DHL was developed and achieved a sensitivity of 100% and a specificity of 97.72%. CONCLUSIONS: The ML and DHL models predicted QA results with a high degree of accuracy. Our predictive QA online platform is offering substantial time savings in terms of accelerator occupancy and working time.

The Impact of Artificial Intelligence CNN Based Denoising on FDG PET Radiomics.

BACKGROUND: With a constantly increasing number of diagnostic images performed each year, Artificial Intelligence (AI) denoising methods offer an opportunity to respond to the growing demand. However, it may affect information in the image in an unknown manner. This study quantifies the effect of AI-based denoising on FDG PET textural information in comparison to a convolution with a standard gaussian postfilter (EARL1). METHODS: The study was carried out on 113 patients who underwent a digital FDG PET/CT (VEREOS, Philips Healthcare). 101 FDG avid lesions were segmented semi-automatically by a nuclear medicine physician. VOIs in the liver and lung as reference organs were contoured. PET textural features were extracted with pyradiomics. Texture features from AI denoised and EARL1 versus original PET images were compared with a Concordance Correlation Coefficient (CCC). Features with CCC values ≥ 0.85 threshold were considered concordant. Scatter plots of variable pairs with R2 coefficients of the more relevant features were computed. A Wilcoxon signed rank test to compare the absolute values between AI denoised and original images was performed. RESULTS: The ratio of concordant features was 90/104 (86.5%) in AI denoised versus 46/104 (44.2%) with EARL1 denoising. In the reference organs, the concordant ratio for AI and EARL1 denoised images was low, respectively 12/104 (11.5%) and 7/104 (6.7%) in the liver, 26/104 (25%) and 24/104 (23.1%) in the lung. SUVpeak was stable after the application of both algorithms in comparison to SUVmax. Scatter plots of variable pairs showed that AI filtering affected more lower versus high intensity regions unlike EARL1 gaussian post filters, affecting both in a similar way. In lesions, the majority of texture features 79/100 (79%) were significantly (p<0.05) different between AI denoised and original PET images. CONCLUSIONS: Applying an AI-based denoising on FDG PET images maintains most of the lesion's texture information in contrast to EARL1-compatible Gaussian filter. Predictive features of a trained model could be thus the same, however with an adapted threshold. Artificial intelligence based denoising in PET is a very promising approach as it adapts the denoising in function of the tissue type, preserving information where it should.

Radioimmunotherapy for Brain Metastases: The Potential for Inflammation as a Target of Choice.

Brain metastases (BM) are frequently detected during the follow-up of patients with malignant tumors, particularly in those with advanced disease. Despite a major progress in systemic anti-cancer treatments, the average overall survival of these patients remains limited (6 months from diagnosis). Also, cognitive decline is regularly reported especially in patients treated with whole brain external beam radiotherapy (WBRT), due to the absorbed radiation dose in healthy brain tissue. New targeted therapies, for an earlier and/or more specific treatment of the tumor and its microenvironment, are needed. Radioimmunotherapy (RIT), a combination of a radionuclide to a specific antibody, appears to be a promising tool. Inflammation, which is involved in multiple steps, including the early phase, of BM development is attractive as a relevant target for RIT. This review will focus on the (1) early biomarkers of inflammation in BM pertinent for RIT, (2) state of the art studies on RIT for BM, and (3) the importance of dosimetry to RIT in BM. These two last points will be addressed in comparison to the conventional EBRT treatment, particularly with respect to the balance between tumor control and healthy tissue complications. Finally, because new diagnostic imaging techniques show a potential for the detection of BM at an early stage of the disease, we focus particularly on this therapeutic window.

The Impact of Resampling and Denoising Deep Learning Algorithms on Radiomics in Brain Metastases MRI.

BACKGROUND: Magnetic resonance imaging (MRI) is predominant in the therapeutic management of cancer patients, unfortunately, patients have to wait a long time to get an appointment for examination. Therefore, new MRI devices include deep-learning (DL) solutions to save acquisition time. However, the impact of these algorithms on intensity and texture parameters has been poorly studied. The aim of this study was to evaluate the impact of resampling and denoising DL models on radiomics. METHODS: Resampling and denoising DL model was developed on 14,243 T1 brain images from 1.5T-MRI. Radiomics were extracted from 40 brain metastases from 11 patients (2049 images). A total of 104 texture features of DL images were compared to original images with paired t-test, Pearson correlation and concordance-correlation-coefficient (CCC). RESULTS: When two times shorter image acquisition shows strong disparities with the originals concerning the radiomics, with significant differences and loss of correlation of 79.81% and 48.08%, respectively. Interestingly, DL models restore textures with 46.15% of unstable parameters and 25.96% of low CCC and without difference for the first-order intensity parameters. CONCLUSIONS: Resampling and denoising DL models reconstruct low resolution and noised MRI images acquired quickly into high quality images. While fast MRI acquisition loses most of the radiomic features, DL models restore these parameters.

[Why proton therapy? And how?] / Apports de la protonthérapie à la radiothérapie d'aujourd'hui, pourquoi, comment ?

Proton therapy is a radiotherapy, based on the use of protons, charged subatomic particles that stop at a given depth depending on their initial energy (pristine Bragg peak), avoiding any output beam, unlike the photons used in most of the other modalities of radiotherapy. Proton therapy has been used for 60 years, but has only become ubiquitous in the last decade because of recent major advances in particle accelerator technology. This article reviews the history of clinical implementation of protons, the nature of the technological advances that now allows its expansion at a lower cost. It also addresses the technical and physical specificities of proton therapy and the clinical situations for which proton therapy may be relevant but requires evidence. Different proton therapy techniques are possible. These are explained in terms of their clinical potential by explaining the current terminology (such as cyclotrons, synchrotrons or synchrocyclotrons, using superconducting magnets, fixed line or arm rotary with passive diffusion delivery or active by scanning) in basic words. The requirements associated with proton therapy are increased due to the precision of the depth dose deposit. The learning curve of proton therapy requires that clinical indications be prioritized according to their associated uncertainties (such as range uncertainties and movement in lung tumors). Many clinical indications potentially fall under proton therapy ultimately. Clinical strategies are explained in a paralleled manuscript.

Uptake of Radium-223 Dichloride and Early [18F]NaF PET Response Are Driven by Baseline [18F]NaF Parameters: a Pilot Study in Castration-Resistant Prostate Cancer Patients.

PURPOSE: The purpose of this study is to identify predictive factors on baseline [18F]NaF positron emission tomography (PET)/computed tomography (CT) of early response to radium-223 dichloride after 3 cycles of treatment in metastatic castration-resistant prostate cancer patients. PROCEDURES: Analysis of 152 metastases was performed in six consecutive patients who underwent [18F]NaF PET/CT at baseline and for early monitoring after 3 cycles of radium-223 dichloride. All metastases depicted on whole-body [18F]NaF PET/CT were contoured and CT (density in Hounsfield units, sclerotic, mixed, or lytic appearance) as well as [18F]NaF [maximum standardized uptake value (SUVmax), SUVmean, and lesion volume (V18F-NaF)] patterns were recorded. Tumor response was defined as percentage change in SUVmax and SUVmean between baseline and post-treatment PET. Bone lesions were defined as stable, responsive, or progressive, according to thresholds derived from a recent multicentre test-retest study in [18F]NaF PET/CT. Total [18F]NaF uptake in metastases, defined as MATV × SUVmean, was correlated to uptake of radium-223 on biodistribution scintigraphy performed 7 days after the first cycle of treatment. RESULTS: Among metastases, 116 involved the axial skeleton and 36 the appendicular skeleton. Lesions were sclerotic in 126 cases and mixed in 26 cases. No lytic lesion was depicted. ROC analysis showed that SUVmax and SUVmean were better predictors of lesion response than V18F-NaF and density on CT (P < 0.0001 and P = 0.001, respectively). SUVmax and SUVmean were predictors of individual tumor response in separate multivariate models (P = 0.01 and P = 0.02, respectively). CT pattern (mixed versus sclerotic) and lesion density were independent predictors only when assessing response with delta SUVmax (P = 0.002 and 0.007, respectively). A good correlation between total [18F]NaF uptake within metastases and their relative radium-223 uptake assessed by two observers 7 days after treatment (r = 0.72 and 0.77, P < 0.0001) was found. CONCLUSIONS: SUVmax and SUVmean on baseline [18F]NaF PET/CT are independent predictors of bone lesions' response to 3 cycles of radium-223 dichloride, supporting the use of NaF to select patients more likely to respond to treatment.

Study of commercial detector responses in non-equilibrium small photon fields of a 1000MU/min CyberKnife system.

PURPOSE: The purpose of this study was to analyze the detector responses in non-equilibrium small photon fields. METHODS: Five detectors (PTW 31014 ionization chamber, PTW 60016, PTW 60017 and Sun Nuclear EDGE diodes and PTW 60003 diamond detector) and one passive dosimeter (Harshaw micro-LiF) as well as a 1000MU/min CyberKnife were modeled with the PENELOPE Monte Carlo code. Field factors, [Formula: see text] were calculated and perturbations due to volume averaging effect, active material effect and coating effect were quantified for the five detectors and passive dosimeter. RESULTS: The PTW 31014 ionization chamber under-response is mainly due to the fluence perturbation caused by the presence of air as detecting material. Regarding diodes, the high density materials used in their active volume and in their coating is responsible for their over-response. Regarding the PTW 60003 diamond, its under-response for the 5mm field size is due to a large volume averaging effect whereas for largest field sizes a nearly perfect compensation between the volume averaging effect and the material effect due to the diamond density occurs. Despite its small size, a volume averaging effect was observed for the micro-LiF for the 5mm field size. CONCLUSION: Perturbations due to volume averaging effect, active material effect and coating effect were investigated and quantified for five active detectors. Since these perturbations can cause opposite effects, wrong conclusions may be drawn regarding the radiological water-equivalence of detectors. Thus, we recommend performing such a study for each novel detector available on the market.

(18)F-FDG Is a Surrogate Marker of Therapy Response and Tumor Recovery after Drug Withdrawal during Treatment with a Dual PI3K/mTOR Inhibitor in a Preclinical Model of Cisplatin-Resistant Ovarian Cancer.

AIM: Targeting the phosphoinositide 3-kinase (PI3K)/mammalian target of rapamycin (mTOR) pathway is a potential means of overcoming chemoresistance in ovarian cancer. We investigated the capability of (18)F-fluororodeoxyglucose ((18)F-FDG) small-animal positron emission tomography (SA-PET) to predict the effects of a dual PI3K/mTOR inhibitor (BEZ-235) in a cisplatin-resistant ovarian cancer model. METHODS: In a first experiment, nude rats bearing subcutaneous SKOV3 tumors received BEZ-235 for 3 days given alone or after paclitaxel and were compared to controls (either untreated or that were given the excipients of paclitaxel and BEZ-235). SA-PET was performed at baseline, on day 3, and day 7. In a second experiment aiming at further exploring the kinetics of (18)F-FDG tumor uptake during the first 48 hours following drug cessation, untreated controls were compared to rats receiving BEZ-235, which were imaged at baseline, on day 3, on day 4, and on day 5. SA-PET results were compared to cell proliferation assessment (Ki-67), PI3K/mTOR downstream target expression studies (pAKT and phospho-eukaryotic translation initiation factor 4E-binding protein 1), and apoptosis evaluation (cleaved caspase-3). RESULTS: In the first experiment, BEZ-235, compared to untreated controls, induced a marked decrease in (18)F-FDG uptake on day 3, which was correlated to a significant decrease in cell proliferation and to a significant PI3K/mTOR pathway inhibition. No tumor necrosis or apoptosis occurred. Four days following treatment cessation, tumor recovery (in terms of PI3K/mTOR inhibition and cell proliferation) occurred and was identified by (18)F-FDG SA-PET. Paclitaxel plus BEZ-235 showed results similar to BEZ-235 alone. In the second experiment, PI3K/mTOR pathways exhibited partial recovery as early as 24 hours following treatment cessation, but both (18)F-FDG SA-PET and cell proliferation remained unchanged. CONCLUSIONS: (18)F-FDG SA-PET is a surrogate marker of target inhibition during treatment with BEZ-235 and predicts tumor recovery 4 days after drug withdrawal, but not during the first 48 hours following drug cessation, when a lag between PI3K/mTOR pathway recovery and metabolic recovery is observed. (18)F-FDG SA-PET could be used for therapy monitoring of PI3K/mTOR inhibitors, but our results also raise questions regarding the potential impact of the delay between PET imaging and the last drug intake on the accuracy of FDG imaging.

Early evaluation of the effects of chemotherapy with longitudinal FDG small-animal PET in human testicular cancer xenografts: early flare response does not reflect refractory disease.

AIM: We aimed to evaluate the usefulness of FDG PET in the early prediction of the effects of chemotherapy on human testicular cancer xenografts. MATERIAL AND METHODS: Nude rats bearing subcutaneous human embryonal carcinoma xenografts received either cisplatin (5 mg/kg) or saline serum. Small-animal PET studies were performed on days 0, 2, 4 and 7 and compared to immunochemistry studies, flow cytometry studies and hexokinase assays. RESULTS: Cisplatin treatment resulted in biphasic FDG uptake evolution: a peak was observed on day 2, followed by a marked decrease on day 7 despite an insignificant change in tumour volume. Similarly, a peak in cyclin A immunostaining was observed on days 2 and 4), followed by a significant decrease on day 7. Flow cytometry showed that the cyclin A peak was not related to increased cell proliferation but was due to a transient S and G(2)/M cell cycle arrest. A marked increase in cell apoptosis was observed from day 2 to day 7. GLUT-1 showed a significant decrease on day 7. Macrophagic infiltrate remained stable except for an increase observed on day 7. In control tumours, continuous growth was observed, all immunostaining markers remaining stable over time. Hexokinase activity was significantly lower on day 7 in treated tumours than in controls. CONCLUSION: FDG PET may be useful in the early evaluation of treatment in patients with testicular cancer. In our model, a very early increased [(18)F]-FDG uptake was related to a transient cell cycle arrest and early stage apoptosis but did not reveal refractory disease.

Gallium-67 scintigraphy in lymphoma: is there a benefit of image fusion with computed tomography?

The usefulness and complementarity of gallium (67Ga) scintigraphy and computed tomography (CT) in the management of patients with lymphoma have been extensively demonstrated. Owing to a lack of anatomical landmarks and physiological distribution of the tracer, precise localisation of abnormalities on 67Ga scintigraphy can be difficult. As fusion imaging techniques between single-photon emission tomography (SPET) and CT have been developed recently, we investigated whether use of CT/67Ga SPET fusion imaging could help in the interpretation of 67Ga scintigraphy. From November 1999 to May 2001, 52 consecutive fusion studies were performed in 38 patients [22 patients with Hodgkin's disease (HD) and 16 patients with non-Hodgkin's lymphoma (NHL)] as part of pre-treatment staging (n=13), treatment evaluation (n=20) or evaluation of suspected recurrence (n=19). 67Ga scintigraphy was carried out 2 and 6 days following the injection of 185-220 MBq 67Ga citrate. On day 2, 67Ga SPET and CT were performed, focussing on the chest and/or the abdomen/pelvis. Data from each imaging method were co-registered using external markers. 67Ga scintigraphy and CT were initially interpreted independently by nuclear medicine physicians and radiologists. CT/67Ga SPET fusion studies were then jointly interpreted and both practitioners indicated when fusion provided additional information in comparison with CT and 67Ga SPET alone. Image fusion was considered to be of benefit in 12/52 (23%) studies which were performed for initial staging (n=4), treatment evaluation (n=4) or evaluation of suspected recurrence (n=4). In these cases, image fusion allowed either confirmation and/or localisation of pathological gallium uptake (n=10) or detection of lesions not visible on CT scan (n=2). Fusion was relevant for discrimination between osseous lesions and lymph node involvement adjacent to bone, especially in the thoracic and lumbar spine and pelvis. In the abdomen and pelvis, fusion helped to differentiate physiological bowel elimination from abnormal uptake, and assisted in precisely locating uptake in neighbouring viscera of the left hypochondrium, including the spleen, left liver lobe, coeliac area, stomach wall and even the splenic flexure. At the thoracic level, fusion also proved useful for demonstrating clearly the relationships of abnormal foci to the pleura, hepatic dome, mediastinum, ribs or thoracic spine. Clinical management was altered by fusion imaging in one patient (chemotherapy was given instead of radiotherapy) and was potentially affected in three other patients (in that, in conjunction with other factors, the results of fusion imaging had an influence on the decision regarding use of irradiation and especially the treatment volume). In conclusion, CT/67Ga SPET fusion imaging allowed precise localisation of gallium uptake and correct attribution to the involved viscera, thereby altering the diagnosis in 20%-25% of studies in comparison with CT and 67Ga SPET analyses alone. CT/67Ga SPET fusion therefore appears valuable in facilitating the interpretation of 67Ga scintigraphy and we recommend its use in patients with lymphoma when CT and 67Ga scintigraphy are planned.