Nivolumab and ipilimumab with concurrent stereotactic radiosurgery for intracranial metastases from non-small cell lung cancer: analysis of the safety cohort for non-randomized, open-label, phase I/II trial

BACKGROUND

Up to 20% of patients with non-small cell lung cancer (NSCLC) develop brain metastasis (BM), for which the current standard of care is radiation therapy with or without surgery. There are no prospective data on the safety of stereotactic radiosurgery (SRS) concurrent with immune checkpoint inhibitor therapy for BM. This is the safety cohort of the phase I/II investigator-initiated trial of SRS with nivolumab and ipilimumab for patients with BM from NSCLC.

PATIENTS AND METHODS

This single-institution study included patients with NSCLC with active BM amenable to SRS. Brain SRS and systemic therapy with nivolumab and ipilimumab were delivered concurrently (within 7 days). The endpoints were safety and 4-month intracranial progression-free survival (PFS).

RESULTS

Thirteen patients were enrolled in the safety cohort, 10 of whom were evaluable for dose-limiting toxicities (DLTs). Median follow-up was 23 months (range 9.7-24.3 months). The median interval between systemic therapy and radiation therapy was 3 days. Only one patient had a DLT; hence, predefined stopping criteria were not met. In addition to the patient with DLT, three patients had treatment-related grade ≥3 adverse events, including elevated liver function tests, fatigue, nausea, adrenal insufficiency, and myocarditis. One patient had a confirmed influenza infection 7 months after initiation of protocol treatment (outside the DLT assessment window), leading to pneumonia and subsequent death from hemophagocytic lymphohistiocytosis. The estimated 4-month intracranial PFS rate was 70.7%.

CONCLUSION

Concurrent brain SRS with nivolumab/ipilimumab was safe for patients with active NSCLC BM. Preliminary analyses of treatment efficacy were encouraging for intracranial treatment response.

Predicting benefit from immune checkpoint inhibitors in patients with non-small-cell lung cancer by CT-based ensemble deep learning: a retrospective study

BACKGROUND

Only around 20-30% of patients with non-small-cell lung cancer (NCSLC) have durable benefit from immune-checkpoint inhibitors. Although tissue-based biomarkers (eg, PD-L1) are limited by suboptimal performance, tissue availability, and tumour heterogeneity, radiographic images might holistically capture the underlying cancer biology. We aimed to investigate the application of deep learning on chest CT scans to derive an imaging signature of response to immune checkpoint inhibitors and evaluate its added value in the clinical context.

METHODS

In this retrospective modelling study, 976 patients with metastatic, EGFR/ALK negative NSCLC treated with immune checkpoint inhibitors at MD Anderson and Stanford were enrolled from Jan 1, 2014, to Feb 29, 2020. We built and tested an ensemble deep learning model on pretreatment CTs (Deep-CT) to predict overall survival and progression-free survival after treatment with immune checkpoint inhibitors. We also evaluated the added predictive value of the Deep-CT model in the context of existing clinicopathological and radiological metrics.

FINDINGS

Our Deep-CT model demonstrated robust stratification of patient survival of the MD Anderson testing set, which was validated in the external Stanford set. The performance of the Deep-CT model remained significant on subgroup analyses stratified by PD-L1, histology, age, sex, and race. In univariate analysis, Deep-CT outperformed the conventional risk factors, including histology, smoking status, and PD-L1 expression, and remained an independent predictor after multivariate adjustment. Integrating the Deep-CT model with conventional risk factors demonstrated significantly improved prediction performance, with overall survival C-index increases from 0·70 (clinical model) to 0·75 (composite model) during testing. On the other hand, the deep learning risk scores correlated with some radiomics features, but radiomics alone could not reach the performance level of deep learning, indicating that the deep learning model effectively captured additional imaging patterns beyond known radiomics features.

INTERPRETATION

This proof-of-concept study shows that automated profiling of radiographic scans through deep learning can provide orthogonal information independent of existing clinicopathological biomarkers, bringing the goal of precision immunotherapy for patients with NSCLC closer.

FUNDING

National Institutes of Health, Mark Foundation Damon Runyon Foundation Physician Scientist Award, MD Anderson Strategic Initiative Development Program, MD Anderson Lung Moon Shot Program, Andrea Mugnaini, and Edward L C Smith.

Integrative Imaging Informatics for Cancer Research: Workflow Automation for Neuro-Oncology (I3CR-WANO)

PURPOSE

Efforts to use growing volumes of clinical imaging data to generate tumor evaluations continue to require significant manual data wrangling, owing to data heterogeneity. Here, we propose an artificial intelligence-based solution for the aggregation and processing of multisequence neuro-oncology MRI data to extract quantitative tumor measurements.

MATERIALS AND METHODS

Our end-to-end framework (1) classifies MRI sequences using an ensemble classifier, (2) preprocesses the data in a reproducible manner, (3) delineates tumor tissue subtypes using convolutional neural networks, and (4) extracts diverse radiomic features. Moreover, it is robust to missing sequences and adopts an expert-in-the-loop approach in which the segmentation results may be manually refined by radiologists. After the implementation of the framework in Docker containers, it was applied to two retrospective glioma data sets collected from the Washington University School of Medicine (WUSM; n = 384) and The University of Texas MD Anderson Cancer Center (MDA; n = 30), comprising preoperative MRI scans from patients with pathologically confirmed gliomas.

RESULTS

The scan-type classifier yielded an accuracy of >99%, correctly identifying sequences from 380 of 384 and 30 of 30 sessions from the WUSM and MDA data sets, respectively. Segmentation performance was quantified using the Dice Similarity Coefficient between the predicted and expert-refined tumor masks. The mean Dice scores were 0.882 (±0.244) and 0.977 (±0.04) for whole-tumor segmentation for WUSM and MDA, respectively.

CONCLUSION

This streamlined framework automatically curated, processed, and segmented raw MRI data of patients with varying grades of gliomas, enabling the curation of large-scale neuro-oncology data sets and demonstrating high potential for integration as an assistive tool in clinical practice.

PocketNet: A Smaller Neural Network for Medical Image Analysis

Medical imaging deep learning models are often large and complex, requiring specialized hardware to train and evaluate these models. To address such issues, we propose the PocketNet paradigm to reduce the size of deep learning models by throttling the growth of the number of channels in convolutional neural networks. We demonstrate that, for a range of segmentation and classification tasks, PocketNet architectures produce results comparable to that of conventional neural networks while reducing the number of parameters by multiple orders of magnitude, using up to 90% less GPU memory, and speeding up training times by up to 40%, thereby allowing such models to be trained and deployed in resource-constrained settings.

Customizable landmark-based field aperture design for automated whole-brain radiotherapy treatment planning

PURPOSE

To develop and evaluate an automated whole-brain radiotherapy (WBRT) treatment planning pipeline with a deep learning-based auto-contouring and customizable landmark-based field aperture design.

METHODS

The pipeline consisted of the following steps: (1) Auto-contour normal structures on computed tomography scans and digitally reconstructed radiographs using deep learning techniques, (2) locate the landmark structures using the beam's-eye-view, (3) generate field apertures based on eight different landmark rules addressing different clinical purposes and physician preferences. Two parallel approaches for generating field apertures were developed for quality control. The performance of the generated field shapes and dose distributions were compared with the original clinical plans. The clinical acceptability of the plans was assessed by five radiation oncologists from four hospitals.

RESULTS

The performance of the generated field apertures was evaluated by the Hausdorff distance (HD) and mean surface distance (MSD) from 182 patients' field apertures used in the clinic. The average HD and MSD for the generated field apertures were 16 ± 7 and 7 ± 3 mm for the first approach, respectively, and 17 ± 7 and 7 ± 3 mm, respectively, for the second approach. The differences regarding HD and MSD between the first and the second approaches were 1 ± 2 and 1 ± 3 mm, respectively. A clinical review of the field aperture design, conducted using 30 patients, achieved a 100% acceptance rate for both the first and second approaches, and the plan review achieved a 100% acceptance rate for the first approach and a 93% acceptance rate for the second approach. The average acceptance rate for meeting lens dosimetric recommendations was 80% (left lens) and 77% (right lens) for the first approach, and 70% (both left and right lenses) for the second approach, compared with 50% (left lens) and 53% (right lens) for the clinical plans.

CONCLUSION

This study provided an automated pipeline with two field aperture generation approaches to automatically generate WBRT treatment plans. Both quantitative and qualitative evaluations demonstrated that our novel pipeline was comparable with the original clinical plans.

Automated Contouring and Planning in Radiation Therapy: What Is 'Clinically Acceptable'?

Developers and users of artificial-intelligence-based tools for automatic contouring and treatment planning in radiotherapy are expected to assess clinical acceptability of these tools. However, what is 'clinical acceptability'? Quantitative and qualitative approaches have been used to assess this ill-defined concept, all of which have advantages and disadvantages or limitations. The approach chosen may depend on the goal of the study as well as on available resources. In this paper, we discuss various aspects of 'clinical acceptability' and how they can move us toward a standard for defining clinical acceptability of new autocontouring and planning tools.

Association between COVID-19 and Postoperative Neurological Complications and Antipsychotic Medication Use after Cancer Surgery: A Retrospective Study

INTRODUCTION

Millions of Americans infected with the severe acute respiratory syndrome-associated coronavirus-19 (COVID-19) need oncologic surgery. Patients with acute or resolved COVID-19 illness complain of neuropsychiatric symptoms. How surgery affects postoperative neuropsychiatric outcomes such as delirium is unknown. We hypothesize that patients with a history of COVID-19 could have an exaggerated risk of developing postoperative delirium after undergoing major elective oncologic surgery.

METHODS

We conducted a retrospective study to determine the association between COVID-19 status and antipsychotic drugs during postsurgical hospitalization as a surrogate of delirium. Secondary outcomes included 30 days of postoperative complications, length of stay, and mortality. Patients were grouped into pre-pandemic non-COVID-19 and COVID-19-positive groups. A 1:2 propensity score matching was used to minimize bias. A multivariable logistic regression model estimated the effects of important covariates on the use of postoperative psychotic medication.

RESULTS

A total of 6003 patients were included in the study. Pre- and post-propensity score matching demonstrated that a history of preoperative COVID-19 did not increase the risk of antipsychotic medications postoperatively. However, respiratory and overall 30-day complications were higher in COVID-19 individuals than in pre-pandemic non-COVID-19 patients. The multivariate analysis showed that the odds of using postoperative antipsychotic medication use for the patients who had COVID-19 compared to those who did not have the infection were not significantly different.

CONCLUSION

A preoperative diagnosis of COVID-19 did not increase the risk of postoperative antipsychotic medication use or neurological complications. More studies are needed to reproduce our results due to the increased concern of neurological events post-COVID-19 infection.

Assessment of intra-fraction motion during frameless image guided Gamma Knife stereotactic radiosurgery

As frameless stereotactic radiosurgery increase in use, the aim of this study was to evaluate intra-fraction motion through cone-beam CT (CBCT) and high-definition motion management (HDMM) systems. Intra-fraction motion measured between localization, repeat localization and post-treatment CBCTs were correlated to intra-faction motion indicated by the HDMM files using the Pearson coefficient (r). A total of 302 plans were reviewed from 263 patients (114 male, 149 female); 216 pairs of localization-repeat localization, and 260 localization-post-treatment CBCTs were analyzed against HDMM logs. We found the magnitude of intra-fraction motion detected by the HDMM system were larger than the corresponding CBCT results.

On the importance of interpretable machine learning predictions to inform clinical decision making in oncology

Machine learning-based tools are capable of guiding individualized clinical management and decision-making by providing predictions of a patient's future health state. Through their ability to model complex nonlinear relationships, ML algorithms can often outperform traditional statistical prediction approaches, but the use of nonlinear functions can mean that ML techniques may also be less interpretable than traditional statistical methodologies. While there are benefits of intrinsic interpretability, many model-agnostic approaches now exist and can provide insight into the way in which ML systems make decisions. In this paper, we describe how different algorithms can be interpreted and introduce some techniques for interpreting complex nonlinear algorithms.

Patient Outcomes in Laparoscopic Appendectomy With Acute Surgical Care Model Compared to Traditional Call

INTRODUCTION

Shift-based models for acute surgical care (ACS), where surgical emergencies are treated by a dedicated team of surgeons working shifts, without a concurrent elective practice, are becoming more common nationwide. We compared the outcomes for appendectomy, one of the most common emergency surgical procedures, between the traditional (TRAD) call and ACS model at the same institution during the same time frame.

METHODS

A retrospective review of patients who underwent laparoscopic appendectomy for acute appendicitis during 2017-2018. ACS and TRAD-patient demographics, clinical presentation, operative details, and outcomes were compared using independent sample t-tests, Wilcoxon rank-sum tests and Fisher's exact or χ² tests. Multiple exploratory regression models were constructed to examine the effects of confounding variables.

RESULTS

Demographics, clinical presentation, and complication rates were similar between groups except for a longer duration of symptoms prior to arrival in the TRAD group (Δ = 0.5 d, P = 0.006). Time from admission to operating room (Δ = -1.85 h, P = 0.003), length of hospital stay (Δ = -2.0 d, P < 0.001), and total cost (Δ = $ -2477.02, P < 0.001) were significantly lower in the ACS group compared to the TRAD group. Furthermore, perforation rates were lower in ACS (8.3% versus 28.6%, P = 0.003). Differences for the outcomes remained significant even after controlling for duration of symptoms prior to arrival (P < 0.05).

CONCLUSIONS

Acute appendicitis managed using the ACS shift-based model seems to be associated with reduced time to operation, hospital stay, and overall cost, with equivalent success rates, compared to TRAD.

A Prospective Study of Conventionally Fractionated Dose Constraints for Reirradiation of Primary Brain Tumors in Adults

PURPOSE

Dose constraints for reirradiation of recurrent primary brain tumors are not well-established. This study was conducted to prospectively evaluate composite dose constraints for conventionally fractionated brain reirradiation.

METHODS AND MATERIALS

A single-institution, prospective study of adults with previously irradiated, recurrent brain tumors was performed. For 95% of patients, electronic dosimetry records from the first course of radiation (RT1) were obtained and deformed onto the simulation computed tomography for the second course of radiation (RT2). Conventionally fractionated treatment plans for RT2 were developed that met protocol-assigned dose constraints for RT2 alone and the composite dose of RT1 + RT2. Prospective composite dose constraints were based on histology, interval since RT1, and concurrent bevacizumab. Patients were followed with magnetic resonance imaging including spectroscopy and perfusion studies. Primary endpoint was the rate of symptomatic brain necrosis at 6 months after RT2.

RESULTS

Patients were enrolled from March 2017 to May 2018; 20 were evaluable. Eighteen had glioma, 1 had atypical choroid plexus papilloma, and 1 had hemangiopericytoma. Nineteen patients were treated with volumetric modulated arc therapy, and one was treated with protons. Median RT1 dose was 57 Gy (range, 50-60 Gy). Median RT1-RT2 interval was 49 months (range, 9-141 months). Median RT2 dose was 42.4 Gy (range, 36-60 Gy). Median planning target volume was 186 cc (range, 8-468 cc). Nineteen of 20 patients (95%) were free of grade 3+ central nervous system necrosis. One patient had grade 3+ necrosis 2 months after RT2; the patient recovered fully and lived another 18 months until dying of disease progression. Median overall survival from RT2 start for all patients was 13.3 months (95% credible interval, 6.3-20.7); for patients with glioblastoma, 11.5 months (95% credible interval, 6.1-20.1).

CONCLUSIONS

Brain reirradiation can be safely performed with conventionally fractionated regimens tailored to previous dose distributions. The prospective composite dose constraints described here are a starting point for future studies of conventionally fractionated reirradiation.

Changes In Outcomes And Factors Associated With Survival In Melanoma Patients With Brain Metastases

BACKGROUD

Treatment options for patients with melanoma brain metastasis (MBM) have changed significantly in the last decade. Few studies have evaluated changes in outcomes and factors associated with survival in MBM patients over time. The aim of this study is to evaluate changes in clinical features and overall survival (OS) for MBM patients.

METHODS

Patients diagnosed with MBMs from 1/1/2009-12/31/2013 (Prior Era; PE) and 1/1/2014-12/31/2018 (Current Era; CE) at The University of Texas MD Anderson Cancer Center were included in this retrospective analysis. The primary outcome measure was OS. Log-rank test assessed differences between groups; multivariable analyses were performed with Cox proportional hazards models and recursive partitioning analysis (RPA).

RESULTS

791 MBM patients (PE, n=332; CE, n=459) were included in analysis. Median OS from MBM diagnosis was 10.3 months (95% CI, 8.9 - 12.4) and improved in the CE versus PE (14.4 vs. 10.3 months, P < .001). Elevated serum LDH was the only factor associated with worse OS in both PE and CE patients. Factors associated with survival in CE MBM patients included patient age, primary tumor Breslow thickness, prior immunotherapy, leptomeningeal disease (LMD), symptomatic MBMs, and whole brain radiation therapy (WBRT). Several factors associated with OS in the PE were not significant in the CE. RPA demonstrated that elevated serum LDH and prior immunotherapy treatment are the most important determinants of survival in CE MBM patients.

CONCLUSIONS

OS and factors associated with OS have changed for MBM patients. This information can inform contemporary patient management and clinical investigations.

Perioperative considerations in patients with chemotherapy-induced cognitive impairment: a narrative review

Patients with cancer may suffer from a decline in their cognitive function after various cancer therapies, including surgery, radiation, and chemotherapy, and in some cases, this decline in cognitive function persists even years after completion of treatment. Chemobrain or chemotherapy-induced cognitive impairment, a well-established clinical syndrome, has become an increasing concern as the number of successfully treated cancer patients has increased significantly. Chemotherapy-induced cognitive impairment can originate from direct neurotoxicity, neuroinflammation, and oxidative stress, resulting in alterations in grey matter volume, white matter integrity, and brain connectivity. Surgery has been associated with exacerbating the inflammatory response associated with chemotherapy and predisposes patients to develop postoperative cognitive dysfunction. As the proportion of patients living longer after these therapies increases, the magnitude of impact and growing concern of post-treatment cognitive dysfunction in these patients has also come to the fore. We review the clinical presentation, potential mechanisms, predisposing factors, diagnostic methods, neuropsychological testing, and imaging findings of chemotherapy-induced cognitive impairment and its intersection with postoperative cognitive dysfunction.

Mathematical modeling of cancer immunotherapy for personalized clinical translation

Encouraging advances are being made in cancer immunotherapy modeling, especially in the key areas of developing personalized treatment strategies based on individual patient parameters, predicting treatment outcomes and optimizing immunotherapy synergy when used in combination with other treatment approaches. Here we present a focused review of the most recent mathematical modeling work on cancer immunotherapy with a focus on clinical translatability. It can be seen that this field is transitioning from pure basic science to applications that can make impactful differences in patients' lives. We discuss how researchers are integrating experimental and clinical data to fully inform models so that they can be applied for clinical predictions, and present the challenges that remain to be overcome if widespread clinical adaptation is to be realized. Lastly, we discuss the most promising future applications and areas that are expected to be the focus of extensive upcoming modeling studies.

Protocol for mathematical prediction of patient response and survival to immune checkpoint inhibitor immunotherapy

This protocol describes the application of a mechanistic mathematical model of immune checkpoint inhibitor (ICI) immunotherapy to patient tumor imaging data for predicting solid tumor response and patient survival under ICI intervention. We describe steps for data collection and processing, data pipelines, and approaches to increase precision. The protocol is highly predictive as early as the first restaging after treatment start and can be used with standard-of-care imaging measures. For complete details on the use and execution of this protocol, please refer to Butner et al. (2020)¹ and Butner et al. (2021).².

NIMG-79. SPATIALLY MAPPED PREDICTIONS OF EVOLVING TUMOR RESPONSE OF HIGH-GRADE GLIOMA VIA IMAGE-DRIVEN MATHEMATICAL MODELING

Timely treatment response assessment of high-grade gliomas (HGG), crucial for driving therapeutic decisions, remains a challenge; as HGGs exhibit variable response to treatment within different sub-regions. Current assessment using multi-parametric MRI (mpMRI) depends largely upon follow-up (FU) imaging timepoints for achieving diagnostic certainty, which delays therapeutic interventions. Mathematical modeling (MM) of tumor growth and treatment response can provide spatiotemporal information of HGG evolution in response to treatment, thus allowing for prospective early identification of resilient tumor subregions. AIMS: We aim to initialize and calibrate an image-driven MM framework to forecast HGG response, both at the end of chemoradiotherapy (CRT) and at 3-month FU.

METHODS

In a prospective clinical study, weekly mpMRIs (post-contrast T1, T2 FLAIR, and diffusion) for patients with HGG receiving CRT were used to describe tumor extent and cellularity. This data collected from baseline (pre-CRT) till week 3 (mid-CRT) was used to calibrate a model family to forecast HGG response for each individual patient at week 6 (end CRT) and at 3-month FU.

RESULTS

Error between the forecasted and observed responses was assessed globally using percent error in tumor volume, and at the local level by Pearson correlation coefficient (PCC). In an initial cohort of 11 patients, our MM framework predictions had a percent error in tumor volume of less than 8.6% and at week 6 RT and less than 20% at 3 months FU. The PCCs were 0.84 at week 6 RT and 0.72 at 3 months FU.

CONCLUSIONS

Temporal consistency across this early evaluation of the model predictions show promise of image-driven MM for HGG response forecasting to guide timely personalized assessment and adjustment of treatment.

NEIM-05 FEASIBILITY OF NAVIGATED TRANSCRANIAL MAGNETIC STIMULATION (NTMS) BASED DIFFUSION TENSOR IMAGING (DTI) TRACTOGRAPHY OF MOTOR PATHWAYS IN PATIENTS UNDERGOING STEREOTACTIC RADIOSURGERY: A CROSS-SECTIONAL DOSIMETRIC AND PATIENT OUTCOMES ANALYSIS

BACKGROUND

No current dose limitations exist for the motor tracts during stereotactic radiosurgery (SRS) planning due to challenges in localizing this region of interest by conventional imaging. Navigated Transcranial Magnetic Stimulation (nTMS) is a non-invasive tool that utilizes electromyographic signal combined with magnetic resonance diffusion tensor imaging (DTI) to functionally map cortical motor tracts. Although nTMS is utilized for functional mapping prior to brain tumor resection, it has not been implemented in SRS planning.

OBJECTIVES

To determine the feasibility of performing nTMS-based DTI in patients treated with SRS and examine the relationship between dose to functionally-defined motor tracts and patient outcomes measured by objective hand function testing and patient reported outcomes (PROs).

METHODS

16 patients treated with SRS to a brain metastasis located near anatomically-defined motor tracts were enrolled on an IRB-approved clinical trial. At median follow-up of 5.4m after SRS, patients underwent nTMS testing, brain MRI with DTI, functional outcomes testing (Pinch Dynamometer,9-Hole Peg Test), and quality-of-life (QOL) PROs (EQ-5D-5L, MDASI-BT). nTMS-seeded DTI tractography was generated (Brainlab iPlan) and imported into GammaPlan for dosimetric evaluation.

RESULTS

Tractography reconstitution was attempted for 8/16 patients and successful in 7/8 (87.5%). One patient who had prior resection of a lesion in the right pre-central gyrus failed to map in the right cortex and was unable to complete functional testing for the affected extremity. Median D_max to the treated motor tracts was 4.6Gy [0.5-13.4Gy]. Median D_mean was 0.9Gy [0-1.2Gy]. Increased D_max correlated with deficits in lateral pinch strength (R²=0.76) and 9-Hole Peg testing time (R²=0.61). Increased D_mean correlated with increased MDASI-BT interference scores (R²=0.93) and EQ5D5L score (R²=0.94) indicating worsened QOL.

CONCLUSIONS

nTMS testing was feasible and dose to nTMS-defined motor tracts correlated with subjective and objective patient outcomes. Future steps will include characterization of motor tract dose tolerance for SRS.

NEIM-06 COMBINING CLINICAL VARIABLES AND RADIOMIC FEATURES TO HELP DISTINGUISH RADIATION NECROSIS FROM TUMOR IN PATIENTS WITH MELANOMA BRAIN METASTASES TREATED WITH RADIOSURGERY

BACKGROUND

Following Gamma Knife SRS (GK-SRS), the conventional imaging characteristics of radiation necrosis (RN) mimic those of tumor progression, introducing considerable uncertainty in diagnosis. Previous studies have identified clinical variables associated with RN; however, diagnosis primarily relied on interpretation of imaging with only a minority confirmed using the gold standard of pathological examination. Furthermore, the cohorts of these studies included a mix of primary histologies.

PURPOSE

To identify the combination of clinical variables and radiomic features most predictive of RN in patients with melanoma brain metastasis (BM) with GK-SRS in order to train a machine learning classifier to distinguish RN from tumor progression.

METHODS

We retrospectively studied 86 patients with a melanoma BM that received initial GK-SRS followed by resection, thereby pathologically confirming tumor or RN. Clinical variables including lesion volume, age at surgery, GK-SRS dose, lesion hemorrhage, lesion location, gender, BM velocity, and drug therapy type were obtained from chart review. We extracted radiomic features from contrast-enhanced T1-weighted MR images using PyRadiomics. A consensus clustering algorithm identified representative radiomic features. Non-parametric hypothesis testing was performed on the clinical variables and representative radiomic features.

RESULTS

Of the 86 patients, 17 (19.8%) patients exhibited RN and 69 exhibited tumor progression. Lesion volume was associated with development of RN (p = 0.038&lt;0.05) with a median volume of 1.5 cc (0.01-26.71 cc). Clustering analysis identified seven representative radiomic features; five were found to have statistically significant association with development of RN.

CONCLUSION

In this dataset with pathologically confirmed diagnoses in a histologically homogeneous patient cohort, we reproduced previously reported findings that the clinical variable of lesion volume is associated with RN and we identified several radiomic features associated with RN in patients with melanoma BM. We are using these variables and features to train a machine learning classifier to distinguish RN from tumor.

Hypothetical generalized framework for a new imaging endpoint of therapeutic activity in early phase clinical trials in brain tumors

Imaging response assessment is a cornerstone of patient care and drug development in oncology. Clinicians/clinical researchers rely on tumor imaging to estimate the impact of new treatments and guide decision making for patients and candidate therapies. This is important in brain cancer, where associations between tumor size/growth and emerging neurological deficits are strong. Accurately measuring the impact of a new therapy on tumor growth early in clinical development, where patient numbers are small, would be valuable for decision making regarding late-stage development activation. Current attempts to measure the impact of a new therapy have limited influence on clinical development, as determination of progression, stability or response does not currently account for individual tumor growth kinetics prior to the initiation of experimental therapies. Therefore, we posit that imaging-based response assessment, often used as a tool for estimating clinical effect, is incomplete as it does not adequately account for growth trajectories or biological characteristics of tumors prior to the introduction of an investigational agent. Here, we propose modifications to the existing framework for evaluating imaging assessment in primary brain tumors that will provide a more reliable understanding of treatment effects. Measuring tumor growth trajectories prior to a given intervention may allow us to more confidently conclude whether there is an anti-tumor effect. This updated approach to imaging-based tumor response assessment is intended to improve our ability to select candidate therapies for later-stage development, including those that may not meet currently sought thresholds for "response" and ultimately lead to identification of effective treatments.

Opportunities for improving brain cancer treatment outcomes through imaging-based mathematical modeling of the delivery of radiotherapy and immunotherapy

Immunotherapy has become a fourth pillar in the treatment of brain tumors and, when combined with radiation therapy, may improve patient outcomes and reduce the neurotoxicity. As with other combination therapies, the identification of a treatment schedule that maximizes the synergistic effect of radiation- and immune-therapy is a fundamental challenge. Mechanism-based mathematical modeling is one promising approach to systematically investigate therapeutic combinations to maximize positive outcomes within a rigorous framework. However, successful clinical translation of model-generated combinations of treatment requires patient-specific data to allow the models to be meaningfully initialized and parameterized. Quantitative imaging techniques have emerged as a promising source of high quality, spatially and temporally resolved data for the development and validation of mathematical models. In this review, we will present approaches to personalize mechanism-based modeling frameworks with patient data, and then discuss how these techniques could be leveraged to improve brain cancer outcomes through patient-specific modeling and optimization of treatment strategies.

Association between telehealth and adherence with patient-reported outcomes (PRO)-based remote symptom monitoring among adolescent/young adults (AYA), middle age, and older adults with cancer

1513

Background: PRO-based remote symptom monitoring favorably impacts quality of life, healthcare utilization, and overall survival in patients (pts) with cancer. However remote PRO completion rates outside of a clinical trial remained widely varied. With the wide adoption of telehealth in cancer care during the pandemic, telehealth’s impact on health behaviors such adherence w remote PROs is not fully characterized. To that end, we investigated PRO completion patterns in routine cancer care, pre- and during the pandemic. Methods: We queried a prospectively maintained institutional database of all PROs remotely delivered to pts at our institution from 1/1/18 to 12/31/21. Pts were divided into 2 time cohorts (“pre-pandemic” 1/1/18 to 3/31/20, “during pandemic” 4/1/20 to 12/31/21) and 3 age cohorts (AYA 15-39y, midage 40-64y, older adults 65y+). We calculated descriptive statistics and compared (t-test, ANOVA) between time and age cohorts and independent variables. Results: Overall 93,875 unique patients over 4 years received 1+ remote PROs as a part of their routine cancer care. PRO response rate increased from 35% prepandemic (12011 of 34742 pts responding) to 67% during pandemic (p <0.00001). To understand patient-level response patterns, we selected one representative global health PRO tool used widely across clinics in our institution and analyzed completion in a representative month over 4 years, 2 before (Oct ’18, ‘19) and 2 mid-pandemic (Oct ’20, ‘21). Overall 2738 pts (median age 60y, range 17-94y; 290 AYA 15-39y, 1444 midage 40-64y, 1004 older adults 65y+) were sent 3249 PROs during these 4m, 1378 PROs to 1075 pts in 2 pre-pandemic months & 1871 to 1663 pts in 2 mid-pandemic months. Overall, PRO response rate increased from 52% pre-pandemic to 81% during, non-responders dropping from 48% to 19%, and response rate without any reminder from the team increasing from 13% pre-pandemic to 79% during. Across all 3 age cohorts, overall PRO response rates increased (AYA up 21%, midage up 27%, seniors up 35%, p 0.012), PRO non-response rate decreased (AYA by 21%, midage by 27%, seniors by 35%, p 0.01), and PRO response rate without reminders from clinic team increased significantly (AYA, by 71%, midage by 78%, senior by 61%, p <0.00001). When further analyzing by visit type during pandemic, the improvements in overall PRO response rates are driven almost exclusively by telehealth where in-person PRO completion decreased by 19% (pre-pandemic 52%, during 33%) while pts who had an upcoming virtual visit had 94% PRO response rate (p < 0.00001). Conclusions: Substantially higher adherence with PRO-based remote symptom monitoring was seen during the pandemic with virtual visits accounting substantially for this broad adherence and the highest increases seen in older adults, highlighting the implications of telehealth on cancer care.

Integrating mechanism-based modeling with biomedical imaging to build practical digital twins for clinical oncology

Digital twins employ mathematical and computational models to virtually represent a physical object (e.g., planes and human organs), predict the behavior of the object, and enable decision-making to optimize the future behavior of the object. While digital twins have been widely used in engineering for decades, their applications to oncology are only just emerging. Due to advances in experimental techniques quantitatively characterizing cancer, as well as advances in the mathematical and computational sciences, the notion of building and applying digital twins to understand tumor dynamics and personalize the care of cancer patients has been increasingly appreciated. In this review, we present the opportunities and challenges of applying digital twins in clinical oncology, with a particular focus on integrating medical imaging with mechanism-based, tissue-scale mathematical modeling. Specifically, we first introduce the general digital twin framework and then illustrate existing applications of image-guided digital twins in healthcare. Next, we detail both the imaging and modeling techniques that provide practical opportunities to build patient-specific digital twins for oncology. We then describe the current challenges and limitations in developing image-guided, mechanism-based digital twins for oncology along with potential solutions. We conclude by outlining five fundamental questions that can serve as a roadmap when designing and building a practical digital twin for oncology and attempt to provide answers for a specific application to brain cancer. We hope that this contribution provides motivation for the imaging science, oncology, and computational communities to develop practical digital twin technologies to improve the care of patients battling cancer.