Machine Learning and Imaging Informatics in Oncology.

In the era of personalized and precision medicine, informatics technologies utilizing machine learning (ML) and quantitative imaging are witnessing a rapidly increasing role in medicine in general and in oncology in particular. This expanding role ranges from computer-aided diagnosis to decision support of treatments with the potential to transform the current landscape of cancer management. In this review, we aim to provide an overview of ML methodologies and imaging informatics techniques and their recent application in modern oncology. We will review example applications of ML in oncology from the literature, identify current challenges and highlight future potentials.

A time- and space-saving Monte Carlo simulation method using post-collimation generative adversarial network for dose calculation of an O-ring gantry Linac.

PURPOSE: This study explores the feasibility of employing Generative Adversarial Networks (GANs) to model the RefleXion X1 Linac. The aim is to investigate the accuracy of dose simulation and assess the potential computational benefits. METHODS: The X1 Linac is a new radiotherapy machine with a binary multi-leaf collimation (MLC) system, facilitating innovative biology-guided radiotherapy. A total of 34 GAN generators, each representing a desired MLC aperture, were developed. Each generator was trained using a phase space file generated underneath the corresponding aperture, enabling the generation of particles and serving as a beam source for Monte Carlo simulation. Dose distributions in water were simulated for each aperture using both the GAN and phase space sources. The agreement between dose distributions was evaluated. The computational time reduction from bypassing the collimation simulation and storage space savings were estimated. RESULTS: The percentage depth dose at 10 cm, penumbra, and full-width half maximum of the GAN simulation agree with the phase space simulation, with differences of 0.4 % ± 0.2 %, 0.32 ± 0.66 mm, and 0.26 ± 0.44 mm, respectively. The gamma passing rate (1 %/1mm) for the planar dose exceeded 90 % for all apertures. The estimated time-saving for simulating an plan using 5766 beamlets was 530 CPU hours. The storage usage was reduced by a factor of 102. CONCLUSION: The utilization of the GAN in simulating the X1 Linac demonstrated remarkable accuracy and efficiency. The reductions in both computational time and storage requirements make this approach highly valuable for future dosimetry studies and beam modeling.

Chest wall pain after single-fraction thoracic stereotactic ablative Radiotherapy: Dosimetric analysis from the iSABR trial.

BACKGROUND AND PURPOSE: Concerns over chest wall toxicity has led to debates on treating tumors adjacent to the chest wall with single-fraction stereotactic ablative radiotherapy (SABR). We performed a secondary analysis of patients treated on the prospective iSABR trial to determine the incidence and grade of chest wall pain and modeled dose-response to guide radiation planning and estimate risk. MATERIALS AND METHODS: This analysis included 99 tumors in 92 patients that were treated with 25 Gy in one fraction on the iSABR trial which individualized dose by tumor size and location. Toxicity events were prospectively collected and graded based on the CTCAE version 4. Dose-response modeling was performed using a logistic model with maximum likelihood method utilized for parameter fitting. RESULTS: There were 22 grade 1 or higher chest wall pain events, including five grade 2 events and zero grade 3 or higher events. The volume receiving at least 11 Gy (V11Gy) and the minimum dose to the hottest 2 cc (D2cc) were most highly correlated with toxicity. When dichotomized by an estimated incidence of ≥ 20 % toxicity, the D2cc > 17 Gy (36.6 % vs. 3.7 %, p < 0.01) and V11Gy > 28 cc (40.0 % vs. 8.1 %, p < 0.01) constraints were predictive of chest wall pain, including among a subset of patients with tumors abutting or adjacent to the chest wall. CONCLUSION: For small, peripheral tumors, single-fraction SABR is associated with modest rates of low-grade chest wall pain. Proximity to the chest wall may not contraindicate single fractionation when using highly conformal, image-guided techniques with sharp dose gradients.

[Diagnostic value of human neutrophil peptide in spontaneous bacterial peritonitis].

OBJECTIVE: To investigate the correlation between human neutrophil peptide (HNP) and spontaneous bacterial peritonitis (SBP) in order to assess the diagnostic value of quantitative measurement of these alpha-defensins. METHODS: Seventy-seven patients with cirrhosis and ascites were divided into two groups according to diagnosis of SBP (n = 45 with SBP and n = 32 without SBP). Twenty-eight healthy individuals were analyzed as controls. HNP was detected by double-antibody sandwich assay. Peripheral white blood cell (WBC) counts, neutrophil ratio, and levels of procalcitonin (PCT) and C-reactive protein (CRP) were determined by standard methods. Receiver operating characteristic (ROC) curves were used to compare the diagnostic values of HNP, PCT and CRP in SBP. RESULTS: There were no significant differences between the three groups (SBP, non-SBP, and healthy controls) for WBC count ((6.01+/-2.25)*109 /L, (4.85+/-1.94)*109 /L, and (5.49+/-1.76)*109 /L; F = 2.91, P more than 0.05) and neutrophil ratio (70.70%+/-10.42%, (68.20%+/-8.97%, and 69.50%+/-8.69%; F = 3.07, P more than 0.05). However, compared to the non-SBP group and the healthy controls, the SBP group showed significantly higher levels of HNP ((9.99+/-3.33) ng/ml and (8.92+/-2.30) ng/ml vs. (17.66+/-6.40) ng/ml; q = 3.20 vs. 3.52, P less than 0.05), serum CRP ((15.08+/-9.95) ng/ml and (5.96+/-2.91) ng/ml vs. (31.32+/-18.65) mg/L; q = 11.03 vs. 3.69, P less than 0.05), and positive rate of PCT (25.0% and 10.0% vs. 62.2%; X2 = 10.41 vs. 15.40, P less than 0.0125). The areas under the curve (AUC) showed the following descending trend: HNP more than PCT more than CRP (0.719, 0.707, and 0.629 respectively). Using cut-off points of 10 ng/ml for HNP, 0.5 ng/ml for PCT, and 8 mg/L for CRP, the respective sensitivities for diagnosis of SBP were 71.1%, 62.2%, and 73.3%, the respective specificities were 71.9%, 75.0%, and 56.3%, and the respective Youden's indexes were 0.430, 0.372, and 0.296. CONCLUSION: HNP is closely related to SBP and can diagnose SBP as reliably as PCT. CRP may help to diagnose SBP, but the results from routine blood testing did not show sufficient statistical significance for diagnosing SBP.

Artificial intelligence (AI) and machine learning (ML) in precision oncology: a review on enhancing discoverability through multiomics integration.

Multiomics data including imaging radiomics and various types of molecular biomarkers have been increasingly investigated for better diagnosis and therapy in the era of precision oncology. Artificial intelligence (AI) including machine learning (ML) and deep learning (DL) techniques combined with the exponential growth of multiomics data may have great potential to revolutionize cancer subtyping, risk stratification, prognostication, prediction and clinical decision-making. In this article, we first present different categories of multiomics data and their roles in diagnosis and therapy. Second, AI-based data fusion methods and modeling methods as well as different validation schemes are illustrated. Third, the applications and examples of multiomics research in oncology are demonstrated. Finally, the challenges regarding the heterogeneity data set, availability of omics data, and validation of the research are discussed. The transition of multiomics research to real clinics still requires consistent efforts in standardizing omics data collection and analysis, building computational infrastructure for data sharing and storing, developing advanced methods to improve data fusion and interpretability, and ultimately, conducting large-scale prospective clinical trials to fill the gap between study findings and clinical benefits.

Interpretable artificial intelligence in radiology and radiation oncology.

Artificial intelligence has been introduced to clinical practice, especially radiology and radiation oncology, from image segmentation, diagnosis, treatment planning and prognosis. It is not only crucial to have an accurate artificial intelligence model, but also to understand the internal logic and gain the trust of the experts. This review is intended to provide some insights into core concepts of the interpretability, the state-of-the-art methods for understanding the machine learning models, the evaluation of these methods, identifying some challenges and limits of them, and gives some examples of medical applications.

3D computational model of oxygen depletion kinetics in brain vasculature during FLASH RT and its implications for in vivo oximetry experiments.

PURPOSE: Ultra-high-dose-rate irradiation, also known as FLASH, has been shown to improve the therapeutic ratio of radiation therapy (RT). The mechanism behind this effect has been partially explained by the radiochemical oxygen depletion (ROD) hypothesis, which attributes the protection of the normal tissue to the induction of transient hypoxia by ROD. To better understand the contribution of oxygen to the FLASH effect, it is necessary to measure oxygen (O2 ) in vivo during FLASH irradiation. This study's goal is to determine the temporal resolution required to accurately measure the rapidly changing oxygen concentration immediately after FLASH irradiation. METHODS: We conducted a computational simulation of oxygen dynamics using a real vascular model that was constructed from a public fluorescence microscopy dataset. The dynamic distribution of oxygen tension (po2 ) during and after FLASH RT was modeled by a partial differential equation (PDE) considering oxygen diffusion, metabolism, and ROD. The underestimation of ROD due to oxygen recovery was evaluated assuming either complete or partial depletion, and a range of possible values for parameters such as oxygen diffusion, consumption, vascular po2 and vessel density. RESULT: The O2 concentration recovers rapidly after FLASH RT. Assuming a temporal resolution of 0.5 s, the estimated ROD is only 50.7% and 36.7% of its actual value in cases of partial and complete depletion, respectively. Additionally, the underestimation of ROD is highly dependent on the vascular density. To estimate ROD rate with 90% accuracy, temporal resolution on the order of milliseconds is required considering the uncertainty in parameters involved, especially, the diverse vascular density of the tissue. CONCLUSION: The rapid recovery of O2 poses a great challenge for in vivo ROD measurements during FLASH RT. Temporal resolution on the order of milliseconds is recommended for ROD measurements in the normal tissue. Further work is warranted to investigate whether the same requirements apply to tumors, given their irregular vasculature.

Artificial Intelligence for Outcome Modeling in Radiotherapy.

Outcome modeling plays an important role in personalizing radiotherapy and finds applications in specialized areas such as adaptive radiotherapy. Conventional outcome models that are based on a simplified understanding of radiobiological effects or empirical fitting often only consider dosimetric information. However, it is recognized that response to radiotherapy is multi-factorial and involves a complex interaction of radiation therapy, patient and treatment factors, and the tumor microenvironment. Recently, large pools of patient-specific biological and imaging data have become available with the development of advanced biotechnology and multi-modality imaging techniques. Given this complexity, artificial intelligence (AI) and machine learning (ML) are valuable to make sense of such a plethora of heterogeneous data and to aid clinicians in their decision-making process. The role of AI/ML has been demonstrated in many retrospective studies and more recently prospective evidence has been emerging as well to support AI/ML for personalized and precision radiotherapy.

Current status and future developments in predicting outcomes in radiation oncology.

Advancements in data-driven technologies and the inclusion of information-rich multiomics features have significantly improved the performance of outcomes modeling in radiation oncology. For this current trend to be sustainable, challenges related to robust data modeling such as small sample size, low size to feature ratio, noisy data, as well as issues related to algorithmic modeling such as complexity, uncertainty, and interpretability, need to be mitigated if not resolved. Emerging computational technologies and new paradigms such as federated learning, human-in-the-loop, quantum computing, and novel interpretability methods show great potential in overcoming these challenges and bridging the gap towards precision outcome modeling in radiotherapy. Examples of these promising technologies will be presented and their potential role in improving outcome modeling will be discussed.

Lung Subregion Partitioning by Incremental Dose Intervals Improves Omics-Based Prediction for Acute Radiation Pneumonitis in Non-Small-Cell Lung Cancer Patients.

Purpose: To evaluate the effectiveness of features obtained from our proposed incremental-dose-interval-based lung subregion segmentation (IDLSS) for predicting grade ≥ 2 acute radiation pneumonitis (ARP) in lung cancer patients upon intensity-modulated radiotherapy (IMRT). (1) Materials and Methods: A total of 126 non-small-cell lung cancer patients treated with IMRT were retrospectively analyzed. Five lung subregions (SRs) were generated by the intersection of the whole lung (WL) and five sub-regions receiving incremental dose intervals. A total of 4610 radiomics features (RF) from pre-treatment planning computed tomographic (CT) and 213 dosiomics features (DF) were extracted. Six feature groups, including WL-RF, WL-DF, SR-RF, SR-DF, and the combined feature sets of WL-RDF and SR-RDF, were generated. Features were selected by using a variance threshold, followed by a Student t-test. Pearson's correlation test was applied to remove redundant features. Subsequently, Ridge regression was adopted to develop six models for ARP using the six feature groups. Thirty iterations of resampling were implemented to assess overall model performance by using the area under the Receiver-Operating-Characteristic curve (AUC), accuracy, precision, recall, and F1-score. (2) Results: The SR-RDF model achieved the best classification performance and provided significantly better predictability than the WL-RDF model in training cohort (Average AUC: 0.98 ± 0.01 vs. 0.90 ± 0.02, p < 0.001) and testing cohort (Average AUC: 0.88 ± 0.05 vs. 0.80 ± 0.04, p < 0.001). Similarly, predictability of the SR-DF model was significantly stronger than that of the WL-DF model in training cohort (Average AUC: 0.88 ± 0.03 vs. 0.70 ± 0.030, p < 0.001) and in testing cohort (Average AUC: 0.74 ± 0.08 vs. 0.65 ± 0.06, p < 0.001). By contrast, the SR-RF model significantly outperformed the WL-RF model only in the training set (Average AUC: 0.93 ± 0.02 vs. 0.85 ± 0.03, p < 0.001), but not in the testing set (Average AUC: 0.79 ± 0.05 vs. 0.77 ± 0.07, p = 0.13). (3) Conclusions: Our results demonstrated that the IDLSS method improved model performance for classifying ARP with grade ≥ 2 when using dosiomics or combined radiomics-dosiomics features.

Integrating Multiomics Information in Deep Learning Architectures for Joint Actuarial Outcome Prediction in Non-Small Cell Lung Cancer Patients After Radiation Therapy.

PURPOSE: Novel actuarial deep learning neural network (ADNN) architectures are proposed for joint prediction of radiation therapy outcomes-radiation pneumonitis (RP) and local control (LC)-in stage III non-small cell lung cancer (NSCLC) patients. Unlike normal tissue complication probability/tumor control probability models that use dosimetric information solely, our proposed models consider complex interactions among multiomics information including positron emission tomography (PET) radiomics, cytokines, and miRNAs. Additional time-to-event information is also used in the actuarial prediction. METHODS AND MATERIALS: Three architectures were investigated: ADNN-DVH considered dosimetric information only; ADNN-com integrated multiomics information; and ADNN-com-joint combined RP2 (RP grade ≥2) and LC prediction. In these architectures, differential dose-volume histograms (DVHs) were fed into 1D convolutional neural networks (CNN) for extracting reduced representations. Variational encoders were used to learn representations of imaging and biological data. Reduced representations were fed into Surv-Nets to predict time-to-event probabilities for RP2 and LC independently and jointly by incorporating time information into designated loss functions. RESULTS: Models were evaluated on 117 retrospective patients and were independently tested on 25 newly accrued patients prospectively. A multi-institutional RTOG0617 data set of 327 patients was used for external validation. ADNN-DVH yielded cross-validated c-indexes (95% confidence intervals) of 0.660 (0.630-0.690) for RP2 prediction and 0.727 (0.700-0.753) for LC prediction, outperforming a generalized Lyman model for RP2 (0.613 [0.583-0.643]) and a generalized log-logistic model for LC (0.569 [0.545-0.594]). The independent internal test and external validation yielded similar results. ADNN-com achieved an even better performance than ADNN-DVH on both cross-validation and independent internal test. Furthermore, ADNN-com-joint, which yielded performance similar to ADNN-com, realized joint prediction with c-indexes of 0.705 (0.676-0.734) for RP2 and 0.740 (0.714-0.765) for LC and achieved an area under a free-response receiving operator characteristic curve (AU-FROC) of 0.729 (0.697-0.773) for the joint prediction of RP2 and LC. CONCLUSION: Novel deep learning architectures that integrate multiomics information outperformed traditional normal tissue complication probability/tumor control probability models in actuarial prediction of RP2 and LC.

Introduction to machine and deep learning for medical physicists.

Recent years have witnessed tremendous growth in the application of machine learning (ML) and deep learning (DL) techniques in medical physics. Embracing the current big data era, medical physicists equipped with these state-of-the-art tools should be able to solve pressing problems in modern radiation oncology. Here, a review of the basic aspects involved in ML/DL model building, including data processing, model training, and validation for medical physics applications is presented and discussed. Machine learning can be categorized based on the underlying task into supervised learning, unsupervised learning, or reinforcement learning; each of these categories has its own input/output dataset characteristics and aims to solve different classes of problems in medical physics ranging from automation of processes to predictive analytics. It is recognized that data size requirements may vary depending on the specific medical physics application and the nature of the algorithms applied. Data processing, which is a crucial step for model stability and precision, should be performed before training the model. Deep learning as a subset of ML is able to learn multilevel representations from raw input data, eliminating the necessity for hand crafted features in classical ML. It can be thought of as an extension of the classical linear models but with multilayer (deep) structures and nonlinear activation functions. The logic of going "deeper" is related to learning complex data structures and its realization has been aided by recent advancements in parallel computing architectures and the development of more robust optimization methods for efficient training of these algorithms. Model validation is an essential part of ML/DL model building. Without it, the model being developed cannot be easily trusted to generalize to unseen data. Whenever applying ML/DL, one should keep in mind, according to Amara's law, that humans may tend to overestimate the ability of a technology in the short term and underestimate its capability in the long term. To establish ML/DL role into standard clinical workflow, models considering balance between accuracy and interpretability should be developed. Machine learning/DL algorithms have potential in numerous radiation oncology applications, including automatizing mundane procedures, improving efficiency and safety of auto-contouring, treatment planning, quality assurance, motion management, and outcome predictions. Medical physicists have been at the frontiers of technology translation into medicine and they ought to be prepared to embrace the inevitable role of ML/DL in the practice of radiation oncology and lead its clinical implementation.

[Clinical implication of expression of perforin/granzyme in peripheral blood lymphocyte in severe hepatitis B].

OBJECTIVE: To explore role of perforin/granzyme expression of peripheral blood lymphocyte in injury of hepatocyte in severe hepatitis B, and evaluate relationship between perforin/granzyme expression levels and hepatitis B virus (HBV)-DNA load. METHODS: Thirty eight patients of severe hepatitis B were enrolled in the study. Fasting venous blood was collected on following morning of admission. Twenty adult healthy subjects served as healthy control group. Perforin/granzyme expression of peripheral blood lymphocyte was detected by flow cytometry, and serum HBV-DNA load was detected by fluorescence quantitative polymerase chain reaction (PCR). RESULTS: Positive rate of perforin/granzyme in peripheral blood lymphocyte in severe hepatitis B was higher than that of the healthy control group [perforin: (43.42+/-19.28)% vs. (19.65+/-9.27)%, granzyme: (40.35+/-12.26)% vs. (22.28+/-9.35)%, both P<0.01]. There was a significant negative correlation between the perforin/granzyme expression of peripheral blood lymphocyte and serum HBV-DNA load (r(perforin) =-0.92, r(granzyme) =-0.96, both P<0.01), the higher serum HBV-DNA load, the lower perforin/granzyme expression in severe hepatitis B. CONCLUSION: Perforin/granzyme overexpression in peripheral blood lymphocyte is an important factor in injury of hepatocyte in patients with severe hepatitis B, and the expression may be involved in HBV-DNA cleanup.

Combining handcrafted features with latent variables in machine learning for prediction of radiation-induced lung damage.

PURPOSE: There has been burgeoning interest in applying machine learning methods for predicting radiotherapy outcomes. However, the imbalanced ratio of a large number of variables to a limited sample size in radiation oncology constitutes a major challenge. Therefore, dimensionality reduction methods can be a key to success. The study investigates and contrasts the application of traditional machine learning methods and deep learning approaches for outcome modeling in radiotherapy. In particular, new joint architectures based on variational autoencoder (VAE) for dimensionality reduction are presented and their application is demonstrated for the prediction of lung radiation pneumonitis (RP) from a large-scale heterogeneous dataset. METHODS: A large-scale heterogeneous dataset containing a pool of 230 variables including clinical factors (e.g., dose, KPS, stage) and biomarkers (e.g., single nucleotide polymorphisms (SNPs), cytokines, and micro-RNAs) in a population of 106 nonsmall cell lung cancer (NSCLC) patients who received radiotherapy was used for modeling RP. Twenty-two patients had grade 2 or higher RP. Four methods were investigated, including feature selection (case A) and feature extraction (case B) with traditional machine learning methods, a VAE-MLP joint architecture (case C) with deep learning and lastly, the combination of feature selection and joint architecture (case D). For feature selection, Random forest (RF), Support Vector Machine (SVM), and multilayer perceptron (MLP) were implemented to select relevant features. Specifically, each method was run for multiple times to rank features within several cross-validated (CV) resampled sets. A collection of ranking lists were then aggregated by top 5% and Kemeny graph methods to identify the final ranking for prediction. A synthetic minority oversampling technique was applied to correct for class imbalance during this process. For deep learning, a VAE-MLP joint architecture where a VAE aimed for dimensionality reduction and an MLP aimed for classification was developed. In this architecture, reconstruction loss and prediction loss were combined into a single loss function to realize simultaneous training and weights were assigned to different classes to mitigate class imbalance. To evaluate the prediction performance and conduct comparisons, the area under receiver operating characteristic curves (AUCs) were performed for nested CVs for both handcrafted feature selections and the deep learning approach. The significance of differences in AUCs was assessed using the DeLong test of U-statistics. RESULTS: An MLP-based method using weight pruning (WP) feature selection yielded the best performance among the different hand-crafted feature selection methods (case A), reaching an AUC of 0.804 (95% CI: 0.761-0.823) with 29 top features. A VAE-MLP joint architecture (case C) achieved a comparable but slightly lower AUC of 0.781 (95% CI: 0.737-0.808) with the size of latent dimension being 2. The combination of handcrafted features (case A) and latent representation (case D) achieved a significant AUC improvement of 0.831 (95% CI: 0.805-0.863) with 22 features (P-value = 0.000642 compared with handcrafted features only (Case A) and P-value = 0.000453 compared to VAE alone (Case C)) with an MLP classifier. CONCLUSION: The potential for combination of traditional machine learning methods and deep learning VAE techniques has been demonstrated for dealing with limited datasets in modeling radiotherapy toxicities. Specifically, latent variables from a VAE-MLP joint architecture are able to complement handcrafted features for the prediction of RP and improve prediction over either method alone.

Artificial Neural Network with Composite Architectures for Prediction of Local Control in Radiotherapy.

In this study, we investigated the application of artificial neural networks (ANNs) with composite architectures into the prediction of local control (LC) of lung cancer patients after radiotherapy. The motivation of this study was to take advantage of the temporal associations among longitudinal (sequential) data to improve the predictive performance of outcome models under the circumstance of limited sample sizes. Two composite architectures: (1) a one dimension (1D) convolutional + fully connected and (2) a locally-connected+ fully connected architectures were implemented for this purpose. Compared with the fully-connected architecture (multi-layer perceptron [MLP]), our composite architectures yielded better predictive performance of LC in lung cancer patients who received radiotherapy. Specifically, in a cohort of 98 patients (29 patients failed locally), the composite architecture of 1D convolutional layers and fully-connected layers achieved an AUC (area under receiver operating characteristic curve) of 0.83 (95% confidence interval (CI): 0.807~0.841) with 18 features (14 features are longitudinal data). Whereas, the composite architecture of locally- connected layers and fully-connected layers achieved an AUC of 0.80 (95%CI: 0.775~0.811). Both outperformed an MLP in the prediction performance with the same set of features, which achieved an AUC of 0.78 (95%CI: 0.751~0.790); (P-values for differences in AUC using the DeLong tests were 1.609 × 10-14and 1.407 × 10-4, respectively).

Balancing accuracy and interpretability of machine learning approaches for radiation treatment outcomes modeling.

Radiation outcomes prediction (ROP) plays an important role in personalized prescription and adaptive radiotherapy. A clinical decision may not only depend on an accurate radiation outcomes' prediction, but also needs to be made based on an informed understanding of the relationship among patients' characteristics, radiation response and treatment plans. As more patients' biophysical information become available, machine learning (ML) techniques will have a great potential for improving ROP. Creating explainable ML methods is an ultimate task for clinical practice but remains a challenging one. Towards complete explainability, the interpretability of ML approaches needs to be first explored. Hence, this review focuses on the application of ML techniques for clinical adoption in radiation oncology by balancing accuracy with interpretability of the predictive model of interest. An ML algorithm can be generally classified into an interpretable (IP) or non-interpretable (NIP) ("black box") technique. While the former may provide a clearer explanation to aid clinical decision-making, its prediction performance is generally outperformed by the latter. Therefore, great efforts and resources have been dedicated towards balancing the accuracy and the interpretability of ML approaches in ROP, but more still needs to be done. In this review, current progress to increase the accuracy for IP ML approaches is introduced, and major trends to improve the interpretability and alleviate the "black box" stigma of ML in radiation outcomes modeling are summarized. Efforts to integrate IP and NIP ML approaches to produce predictive models with higher accuracy and interpretability for ROP are also discussed.

Deep reinforcement learning for automated radiation adaptation in lung cancer.

PURPOSE: To investigate deep reinforcement learning (DRL) based on historical treatment plans for developing automated radiation adaptation protocols for nonsmall cell lung cancer (NSCLC) patients that aim to maximize tumor local control at reduced rates of radiation pneumonitis grade 2 (RP2). METHODS: In a retrospective population of 114 NSCLC patients who received radiotherapy, a three-component neural networks framework was developed for deep reinforcement learning (DRL) of dose fractionation adaptation. Large-scale patient characteristics included clinical, genetic, and imaging radiomics features in addition to tumor and lung dosimetric variables. First, a generative adversarial network (GAN) was employed to learn patient population characteristics necessary for DRL training from a relatively limited sample size. Second, a radiotherapy artificial environment (RAE) was reconstructed by a deep neural network (DNN) utilizing both original and synthetic data (by GAN) to estimate the transition probabilities for adaptation of personalized radiotherapy patients' treatment courses. Third, a deep Q-network (DQN) was applied to the RAE for choosing the optimal dose in a response-adapted treatment setting. This multicomponent reinforcement learning approach was benchmarked against real clinical decisions that were applied in an adaptive dose escalation clinical protocol. In which, 34 patients were treated based on avid PET signal in the tumor and constrained by a 17.2% normal tissue complication probability (NTCP) limit for RP2. The uncomplicated cure probability (P+) was used as a baseline reward function in the DRL. RESULTS: Taking our adaptive dose escalation protocol as a blueprint for the proposed DRL (GAN + RAE + DQN) architecture, we obtained an automated dose adaptation estimate for use at ∼2/3 of the way into the radiotherapy treatment course. By letting the DQN component freely control the estimated adaptive dose per fraction (ranging from 1-5 Gy), the DRL automatically favored dose escalation/de-escalation between 1.5 and 3.8 Gy, a range similar to that used in the clinical protocol. The same DQN yielded two patterns of dose escalation for the 34 test patients, but with different reward variants. First, using the baseline P+ reward function, individual adaptive fraction doses of the DQN had similar tendencies to the clinical data with an RMSE = 0.76 Gy; but adaptations suggested by the DQN were generally lower in magnitude (less aggressive). Second, by adjusting the P+ reward function with higher emphasis on mitigating local failure, better matching of doses between the DQN and the clinical protocol was achieved with an RMSE = 0.5 Gy. Moreover, the decisions selected by the DQN seemed to have better concordance with patients eventual outcomes. In comparison, the traditional temporal difference (TD) algorithm for reinforcement learning yielded an RMSE = 3.3 Gy due to numerical instabilities and lack of sufficient learning. CONCLUSION: We demonstrated that automated dose adaptation by DRL is a feasible and a promising approach for achieving similar results to those chosen by clinicians. The process may require customization of the reward function if individual cases were to be considered. However, development of this framework into a fully credible autonomous system for clinical decision support would require further validation on larger multi-institutional datasets.

[Anti-HBV efficacy of bifendate in treatment of chronic hepatitis B, a primary study].

OBJECTIVE: To investigate the anti-HBV efficacy of bifendate in treatment of chronic hepatitis B. METHODS: A total of 119 patients with chronic hepatitis B were randomly divided into treatment group (n = 65, aged 24 +/- 12) and control group (n = 54, aged 25 +/- 11). In the treatment group every patient was given higher doses bifendate pills ( 12 age, 45 approximately 67.5 mg/d) for up to 12 months. Hepatic function test was performed and HBeAg, HBeAb and HBV DNA were detected at regular intervals in all patients. RESULTS: The serum alanine amonotransferase (ALT) decreased to normal only one month later in 70.76% of patients in the treatment group and decreased to normal at least in 2 approximately 3 months in the control group (P < 0.01). The serum conversion rates of HBeAg, HBeAb and HBV DNA in the treatment group were 44.4%, 29.3%, 38.5%, respectively, which were significantly higher than those in the control group (P < 0.01). No noticeable side effect was observed. CONCLUSION: Higher doses of bifendate taken for a long term has remarkable anti-HBV efficacy in treatment of chronic hepatitis B.