DASS Good: Explainable Data Mining of Spatial Cohort Data.

Developing applicable clinical machine learning models is a difficult task when the data includes spatial information, for example, radiation dose distributions across adjacent organs at risk. We describe the co-design of a modeling system, DASS, to support the hybrid human-machine development and validation of predictive models for estimating long-term toxicities related to radiotherapy doses in head and neck cancer patients. Developed in collaboration with domain experts in oncology and data mining, DASS incorporates human-in-the-loop visual steering, spatial data, and explainable AI to augment domain knowledge with automatic data mining. We demonstrate DASS with the development of two practical clinical stratification models and report feedback from domain experts. Finally, we describe the design lessons learned from this collaborative experience.

The PRO-ACTIVE trial protocol: a randomized study comparing the effectiveness of PROphylACTic swallow InterVEntion for patients receiving radiotherapy for head and neck cancer.

BACKGROUND: Swallowing therapy is commonly provided as a treatment to lessen the risk or severity of dysphagia secondary to radiotherapy (RT) for head and neck cancer (HNC); however, best practice is not yet established. This trial will compare the effectiveness of prophylactic (high and low intensity) versus reactive interventions for swallowing in patients with HNC undergoing RT. METHODS: This multi-site, international randomized clinical trial (RCT) will include 952 adult patients receiving radiotherapy for HNC and who are at high risk for post-RT dysphagia. Participants will be randomized to receive one of three interventions for swallowing during RT: RE-ACTIVE, started promptly if/when dysphagia is identified; PRO-ACTIVE EAT, low intensity prophylactic intervention started before RT commences; or, PRO-ACTIVE EAT+EXERCISE, high intensity prophylactic intervention also started before RT commences. We hypothesize that the PRO-ACTIVE therapies are more effective than late RE-ACTIVE therapy; and, that the more intensive PRO-ACTIVE (EAT + EXERCISE) is superior to the low intensive PRO-ACTIVE (EAT). The primary endpoint of effectiveness is duration of feeding tube dependency one year post radiation therapy, selected as a pragmatic outcome valued equally by diverse stakeholders (e.g., patients, caregivers and clinicians). Secondary outcomes will include objective measures of swallow physiology and function, pneumonia and weight loss, along with various patient-reported swallowing-related outcomes, such as quality of life, symptom burden, and self-efficacy. DISCUSSION: Dysphagia is a common and potentially life-threatening chronic toxicity of radiotherapy, and a priority issue for HNC survivors. Yet, the optimal timing and intensity of swallowing therapy provided by a speech-language pathologist is not known. With no clearly preferred strategy, current practice is fraught with substantial variation. The pragmatic PRO-ACTIVE trial aims to specifically address the decisional dilemma of when swallowing therapy should begin (i.e., before or after a swallowing problem develops). The critical impact of this dilemma is heightened by the growing number of young HNC patients in healthcare systems that need to allocate resources most effectively. The results of the PRO-ACTIVE trial will address the global uncertainty regarding best practice for dysphagia management in HNC patients receiving radiotherapy. TRIAL REGISTRATION: The protocol is registered with the US Patient Centered Outcomes Research Institute, and the PRO-ACTIVE trial was prospectively registered at ClinicalTrials.gov , under the identifier NCT03455608 ; First posted: Mar 6, 2018; Last verified: Jun 17, 2021. Protocol Version: 1.3 (January 27, 2020).

CD8 infiltration is associated with disease control and tobacco exposure in intermediate-risk oropharyngeal cancer.

Oropharyngeal squamous cell carcinoma (OPSCC) incidence is increasing at a nearly epidemic rate, largely driven by the human papillomavirus (HPV). Despite the generally favorable clinical outcomes of patients with HPV driven (HPV+) OPSCC, a significant subset of HPV tumors associated with tobacco exposure have diminished treatment response and worse survival. The tumor immune microenvironment (TIME) has been shown to be a critical driver of treatment response and oncologic outcomes in OPSCC generally and HPV+ OPSCC more specifically. However, the impact of tobacco exposure on the TIME in OPSCC patients remains unclear. We analyzed the relationship between TIME, tobacco exposure and clinical outcomes in OPSCC patients (n = 143) with extensive tobacco exposure (median pack-years = 40). P16 overexpression, a surrogate marker of HPV association, was a strong predictor of relapse-free (RFS) and overall survival (OS) (p < 0.001, p < 0.001 respectively) regardless of tobacco exposure and associated strongly with differential infiltration of the tumor by both CD3 and CD8 lymphocytes measured via immunohistochemistry (p < 001, p < 0.001 respectively). CD3 and CD8 infiltration was a strong predictor of RFS and OS and associated strongly with disease stage (AJCC 8th Edition Staging Manual). Tobacco exposure correlated significantly (p < 0.001) with decreased CD8 infiltration in p16+ OPSCC tumors. Our findings demonstrate that the HPV+ OPSCC clinical outcomes are strongly correlated with the TIME, which is potentially modulated by tobacco exposure. Immunomodulatory strategies targeting this disease in smokers must take into consideration the potential modifying effects of tobacco exposure on treatment effectiveness and clinical outcomes.

A spatial neighborhood methodology for computing and analyzing lymph node carcinoma similarity in precision medicine.

Precision medicine seeks to tailor therapy to the individual patient, based on statistical correlates from patients who are similar to the one under consideration. These correlates can and should go beyond genetics, and in general, beyond tabular or array data that can be easily represented computationally and compared. For example, in many types of cancer, cancer treatment and toxicity depend in large measure on the spatial disease spread-e.g., metastasizes to regional lymph nodes in head and neck cancer. However, there is currently a lack of methodology for integrating spatial information when considering patient similarity. We present a novel modeling methodology for the comparison of cancer patients within a cohort, based on the spatial spread of the lymph nodes affected in each patient. The method uses a topological map, bigrams, and hierarchical clustering to group patients based on their similarity. We compare this approach against a nonspatial (categorical) similarity approach where patients are binned solely by their affected nodes. We present similarity results on a 582 head and neck cancer patient cohort, along with two visual abstractions for analysis of the results, and we present clinician feedback. Our novel methodology partitions a patient cohort into clinically meaningful groups more susceptible to treatment side-effects. Such spatially-aware similarity approaches can help maximize the effectiveness of each patient's treatment.

Cohort-based T-SSIM Visual Computing for Radiation Therapy Prediction and Exploration.

We describe a visual computing approach to radiation therapy (RT) planning, based on spatial similarity within a patient cohort. In radiotherapy for head and neck cancer treatment, dosage to organs at risk surrounding a tumor is a large cause of treatment toxicity. Along with the availability of patient repositories, this situation has lead to clinician interest in understanding and predicting RT outcomes based on previously treated similar patients. To enable this type of analysis, we introduce a novel topology-based spatial similarity measure, T-SSIM, and a predictive algorithm based on this similarity measure. We couple the algorithm with a visual steering interface that intertwines visual encodings for the spatial data and statistical results, including a novel parallel-marker encoding that is spatially aware. We report quantitative results on a cohort of 165 patients, as well as a qualitative evaluation with domain experts in radiation oncology, data management, biostatistics, and medical imaging, who are collaborating remotely.

MRIgRT head and neck anthropomorphic QA phantom: Design, development, reproducibility, and feasibility study.

PURPOSE: The purpose of this paper was to design, manufacture, and evaluate a tissue equivalent, dual magnetic resonance/computed tomography (MR/CT) visible anthropomorphic head and neck (H&N) phantom. This phantom was specially designed as an end-to-end quality assurance (QA) tool for MR imaging guided radiotherapy (MRIgRT) systems participating in NCI-sponsored clinical trials. METHOD: The MRIgRT H&N phantom was constructed using a water-fillable acrylic shell and a custom insert that mimics an organ at risk (OAR) and target structures. The insert consists of a primary and secondary planning target volume (PTV) manufactured of a synthetic Clear Ballistic gel, an acrylic OAR and surrounding tissue fabricated using melted Superflab. Radiochromic EBT3 film and thermoluminescent detectors (TLDs) were used to measure the dose distribution and absolute dose, respectively. The phantom was evaluated by conducting an end-to-end test that included: imaging on a GE Lightspeed CT simulator, planning on Monaco treatment planning software (TPS), verifying treatment setup with MR, and irradiating on Elekta's 1.5 T Unity MR linac system. The phantom was irradiated three times using the same plan to determine reproducibility. Three institutions, equipped with either ViewRay MRIdian 60 Co or ViewRay MRIdian Linac, were used to conduct a feasibility study by performing independent end-to-end studies. Thermoluminescent detectors were evaluated in both reproducibility and feasibility studies by comparing ratios of measured TLD to reported TPS calculated values. Radiochromic film was used to compare measured planar dose distributions to expected TPS distributions. Film was evaluated by using an in-house gamma analysis software to measure the discrepancies between film and TPS. RESULTS: The MRIgRT H&N phantom on the Unity system resulted in reproducible TLD doses (SD < 1.5%). The measured TLD to calculated dose ratios for the Unity system ranged from 0.94 to 0.98. The Viewray dose result comparisons had a larger range (0.95-1.03) but these depended on the TPS dose calculations from each site. Using a 7%/4 mm gamma analysis, Viewray institutions had average axial and sagittal passing rates of 97.3% and 96.2% and the Unity system had average passing rates of 97.8% and 89.7%, respectively. All of the results were within the Imaging and Radiation Oncology Core in Houston (IROC-Houston) standard credentialing criteria of 7% on TLDs, and >85% of pixels passing gamma analysis using 7%/4 mm on films. CONCLUSIONS: An MRIgRT H&N phantom that is tissue equivalent and visible on both CT and MR was developed. The results from initial reproducibility and feasibility testing of the MRIgRT H&N phantom using the tested MGIgRT systems suggests the phantom's potential utility as a credentialing tool for NCI-clinical trials.

Geometric and dosimetric evaluations of atlas-based segmentation methods of MR images in the head and neck region.

Owing to its excellent soft-tissue contrast, magnetic resonance (MR) imaging has found an increased application in radiation therapy (RT). By harnessing these properties for treatment planning, automated segmentation methods can alleviate the manual workload burden to the clinical workflow. We investigated atlas-based segmentation methods of organs at risk (OARs) in the head and neck (H&N) region using one approach that selected the most similar atlas from a library of segmented images and two multi-atlas approaches. The latter were based on weighted majority voting and an iterative atlas-fusion approach called STEPS. We built the atlas library from pre-treatment T1-weighted MR images of 12 patients with manual contours of the parotids, spinal cord and mandible, delineated by a clinician. Following a leave-one-out cross-validation strategy, we measured the geometric accuracy by calculating Dice similarity coefficients (DSC), standard and 95% Hausdorff distances (HD and HD95), and the mean surface distance (MSD), whereby the manual contours served as the gold standard. To benchmark the algorithm, we determined the inter-observer variability (IOV) between three observers. To investigate the dosimetric effect of segmentation inaccuracies, we implemented an auto-planning strategy within the treatment planning system Monaco (Elekta AB, Stockholm, Sweden). For each set of auto-segmented OARs, we generated a plan for a 9-beam step and shoot intensity modulated RT treatment, designed according to our institution's clinical H&N protocol. Superimposing the dose distributions on the gold standard OARs, we calculated dose differences to OARs caused by delineation differences between auto-segmented and gold standard OARs. We investigated the correlations between geometric and dosimetric differences. The mean DSC was larger than 0.8 and the mean MSD smaller than 2 mm for the multi-atlas approaches, resulting in a geometric accuracy comparable to previously published results and within the range of the IOV. While dosimetric differences could be as large as 23% of the clinical goal, treatment plans fulfilled all imposed clinical goals for the gold standard OARs. Correlations between geometric and dosimetric measures were low with R2 < 0.5. The geometric accuracy and the ability to achieve clinically acceptable treatment plans indicate the suitability of using atlas-based contours for RT treatment planning purposes. The low correlations between geometric and dosimetric measures suggest that geometric measures alone are not sufficient to predict the dosimetric impact of segmentation inaccuracies on treatment planning for the data utilised in this study.

The role of radiation therapy in benign diseases.

Although adequate prospective data are lacking, radiation therapy seems to be effective for many benign diseases and remains one of the treatment modalities in the armamentarium of medical professionals. Just as medication has potential adverse effects, and surgery has attendant morbidity, irradiation sometimes can be associated with acute and chronic sequelae. In selecting the mode of treatment, most radiation oncologists consider the particular problem to be addressed and the goal of therapy in the individual patient. It is the careful and judicial use of any therapy that identifies the professional. With an understanding of the current clinical data, treatment techniques, cost, and potential detriment, the goal is to provide long-term control of the disease while minimizing unnecessary treatment and potential risks of side effects. The art lies in balancing benefits against risks.

Time and PSA threshold model prognosticates long-term overall and disease-specific survival in prostate cancer patients as early as 3 months after external beam radiation therapy.

The specific aim of this analysis was to evaluate the capability of a time and prostate-specific antigen (PSA) threshold model to prognosticate overall survival (OS) and disease-specific survival (DSS) based on early PSA kinetics after radiotherapy for prostate cancer by retrospective review of outcomes in 918 patients. Crossing below analyzed PSA thresholds at specific defined time points reduced disease-specific death hazard ratios to relative to the cohort above threshold. The time and PSA threshold model demonstrates the ability to prognosticate OS and DSS as early as 3 months post-radiotherapy for prostate cancer.

A reliable assessment of 8-oxo-2-deoxyguanosine levels in nuclear and mitochondrial DNA using the sodium iodide method to isolate DNA.

A major controversy in the area of DNA biochemistry concerns the actual in vivo levels of oxidative damage in DNA. We show here that 8-oxo-2-deoxyguanosine (oxo8dG) generation during DNA isolation is eliminated using the sodium iodide (NaI) isolation method and that the level of oxo8dG in nuclear DNA (nDNA) is almost one-hundredth of the level obtained using the classical phenol method. We found using NaI that the ratio of oxo8dG/10(5 )deoxyguanosine (dG) in nDNA isolated from mouse tissues ranged from 0.032 +/- 0.002 for liver to 0.015 +/- 0.003 for brain. We observed a significant increase (10-fold) in oxo8dG in nDNA isolated from liver tissue after 2 Gy of gamma-irradiation when NaI was used to isolate DNA. The turnover of oxo8dG in nDNA was rapid, e.g. disappearance of oxo8dG in the mouse liver in vivo after gamma-irradiation had a half-life of 11 min. The levels of oxo8dG in mitochondrial DNA isolated from liver, heart and brain were 6-, 16- and 23-fold higher than nDNA from these tissues. Thus, our results showed that the steady-state levels of oxo8dG in mouse tissues range from 180 to 360 lesions in the nuclear genome and from one to two lesions in 100 mitochondrial genomes.