Patient-specific Treatment Plans Research Articles

Are Thinking Machines Breaking New Frontiers in Neuro-Oncology? A Narrative Review on the Emerging Role of Machine Learning in Neuro-Oncological Practice.

Medical science in general and oncology in particular are dynamic, rapidly evolving subjects. Brain and spine tumors, whether primary or secondary, constitute a significant number of cases in any oncological practice. With the rapid influx of data in all aspects of neuro-oncological care, it is almost impossible for practicing clinicians to remain abreast with the current trends, or to synthesize the available data for it to be maximally beneficial for their patients. Machine-learning (ML) tools are fast gaining acceptance as an alternative to conventional reliance on online data. ML uses artificial intelligence to provide a computer algorithm-based information to clinicians. Different ML models have been proposed in the literature with a variable degree of precision and database requirements. ML can potentially solve the aforementioned problems for practicing clinicians by not just extracting and analyzing useful data, by minimizing or eliminating certain potential areas of human error, by creating patient-specific treatment plans, and also by predicting outcomes with reasonable accuracy. Current information on ML in neuro-oncology is scattered, and this literature review is an attempt to consolidate it and provide recent updates.

How can scanned proton beam treatment planning for low-grade glioma cope with increased distal RBE and locally increased radiosensitivity for late MR-detected brain lesions?

A novel risk model has recently been proposed for the occurrence of late contrast-enhancing brain lesions (CEBLs) after proton irradiation of low-grade glioma (LGG) patients. It predicts a strong dependence on dose-weighted linear-energy transfer (LETd effect) and an increased radiosensitivity of the ventricular proximity, a 4-mm fringe surrounding the ventricular system (VP4mm effect). On this basis, we investigated (A) how these two risk factors and patient-specific anatomical and treatment plan (TP) features contribute to normal tissue complication probability (NTCP) and (B) if conventional LETd -reduction techniques like multiple-field TP are able to reduce NTCP. (A) The LGG model cohort (N=110) was stratified with respect to prescribed dose, tumor grade, and treatment field configuration. NTCP predictions and CEBL occurrence rates per strata were analyzed. (B) The effect of multiple-field TP was investigated in two patient groups: (i) nine high-risk subjects with extended lateral target volumes who had developed CEBLs after single-beam treatments were retrospectively replanned with a clinical standard two-field setting using almost orthogonal fields and strictly opposing fields, (ii) single-field treatments were simulated for seven low-risk patients with small central target volumes clinically treated with two strictly opposing fields. (A) In the model cohort, we identified the exposure of the radiosensitive VP4mm fringe with proton field components of increased biological effectiveness as dominant NTCP driving factor. We observed that larger target volumes and location lateral to the main ventricles, both being characteristic for WHO°II tumors, presented with the highest complication risks. Among subjects of an equal dose prescription of 54Gy(RBE), the highest median NTCP was obtained for the WHO°II group treated with two fields using sharp angles. (B) Regarding the effect of multiple-field plans, we found that an NTCP reduction was only achievable in the low-risk group where the LETd effect dominates and the VP4mm effect is small. NTCP of the single-field plans was 23% higher compared to the clinical opposing field plan. In the high-risk group, where the VP4mm effect dominates the risk, both two-field scenarios yielded 44% higher NTCP predictions compared to the clinical single-field plans. The interplay of an increased radiosensitivity in the VP4mm fringe with proton field components of increased biological effectiveness creates a geometric complexity that can hardly be managed by current clinical TP. Our results underline that advanced biologically guided TP approaches become crucial for an effective risk minimization in proton therapy of LGG.

Editorial: Interstitial Lung Disease in the Context of Systemic Disease: Pathophysiology, Treatment and Outcomes

Interstitial lung disease (ILD) is an umbrella term for many different disease entities causing inflammation and fibrosis of the lung parenchyma. These can be broadly divided into five categories based on etiology (1): (1) ILD related to a distinct primary disease (e.g., sarcoidosis), (2) ILD related to environmental factors (e. g., hypersensitivity pneumonitis), (3) ILD induced by drugs or irradiation, (4) idiopathic interstitial pneumonias (e.g., idiopathic pulmonary fibrosis), and (5) ILD related to connective tissue diseases (CTD) (1). While all these entities require thorough and multidisciplinary assessment to ascertain a diagnosis, establish the need for diagnostic procedures, and recommend a patient-specific treatment plan, ILDs associated with systemic diseases are particularly challenging. In many cases, optimal treatment for involvement of other organ systems needs to be balanced with the choice of ILD-directed therapies. For the highly heterogeneous group of patients who cannot be given a definite diagnosis of an autoimmune rheumatic disease but who demonstrate certain clinical, radiographic, and/or serological features suggestive of a CTD, the term interstitial pneumonia with autoimmune features (IPAF) has been coined (2), but its clinical value remains to be defined. Due to its complexity, management of ILDs associated with systemic disease requires multidisciplinary care, including pulmonologists, rheumatologists, and radiologists, often with critical input from other specialties, such as pathologists, dermatologists, or neurologists. In this Research Topic, we hope to present novel research and state-of-the-art reviews as relevant to the care of patients with ILD and systemic diseases.

Platelet Dysfunction after Traumatic Brain Injury: A Review.

Coagulopathy is a known sequela of traumatic brain injury (TBI) and can lead to increased morbidity and mortality. Platelet dysfunction has been identified as one of several etiologies of coagulopathy following TBI and has been associated with poor outcomes. Regardless of whether the platelet dysfunction occurs as a direct consequence of the injury or because of pre-existing medical comorbidities or medication use, accurate detection and monitoring of response to therapy is key to optimal patient care. Platelet transfusion has been proposed as a potential therapeutic intervention to treat platelet dysfunction, with several studies using platelet function assays to monitor response. The development of increasingly precise diagnostic testing is providing enhanced understanding of the specific derangement in the hemostatic process, allowing clinicians to provide patient-specific treatment plans. There is wide variability in the currently available literature on the incidence and clinical significance of platelet dysfunction following TBI, which creates challenges with developing evidence-based management guidelines. The relatively high prevalence of platelet inhibitor therapy serves as an additional confounding factor. In addition, the data are largely retrospective in nature. We performed a literature review to provide clarity on this clinical issue. We reviewed 348 abstracts, and included 97 manuscripts in our final literature review. Based on the currently available research, platelet dysfunction has been consistently demonstrated in patients with moderate-severe TBI. We recommend the use of platelet functional assays to evaluate patients with TBI. Platelet transfusion directed at platelet dysfunction may lead to improved clinical outcome. A randomized trial guided by implementation science could improve the applicability of these practices.

Fast Optimization of Temperature Focusing in Hyperthermia Treatment of Sub-Superficial Tumors

Microwave hyperthermia aims at selectively heating cancer cells to a supra-physiological temperature. For non-superficial tumors, this can be achieved by means of an antenna array equipped with a proper cooling system (the water bolus) to avoid overheating of the skin. In patient-specific treatment planning, antenna feedings are optimized to maximize the specific absorption rate (SAR) inside the tumor, or to directly maximize the temperature there, involving a higher numerical cost. We present here a method to effect a low-complexity temperature-based planning. It arises from recognizing that SAR and temperature have shifted peaks due to thermal boundary conditions at the water bolus and for physiological effects like air flow in respiratory ducts. In our method, temperature focusing on the tumor is achieved via a SAR-based optimization of the antenna excitations, but optimizing its target to account for the cooling effects. The temperature optimization process is turned into finding a SAR peak position that maximizes the chosen temperature objective function. Application of this method to the 3D head and neck region provides a temperature coverage that is consistently better than that obtained with SAR-optimization alone, also considering uncertainties in thermal parameters. This improvement is obtained by solving the bioheat equation a reduced number of times, avoiding its inclusion in a global optimization process.

Motion laboratory gait analysis and orthopedic resident education: preliminary results.

Gait analysis in the pediatric and adult orthopedic patient populations can adjunct the diagnosis and treatment of a multitude of musculoskeletal conditions. Understanding of normal and abnormal gait biomechanics is an important aspect of orthopedic residency; yet, there is great variability in the time residency programs dedicate to gait analysis education. The purpose of this study was to investigate if formal gait analysis education during residency improves an orthopedic resident's understanding of normal and pathologic gait. Five residency programs consisting of 81 resident subjects were surveyed at the beginning of the 2016-2017 academic year. The residents were divided into those with formal gait analysis education (group A) and those without (group B). Each resident was sent an online survey with 11 questions from former orthopedic in-training exams (OITE) regarding gait pattern analysis. The average number of correct questions was compared between the two groups with Student's t-test. Fifty-three of the 81 surveys sent out were completed. There were 23 subjects in group A and 30 in group B. All five programs and all postgraduate years (PGYs) were represented (PGY1: 10, PGY2: 12, PGY3: 12, PGY4: 6, PGY5 12). The average score for all residents was 5.6 out of 11 correct (51%). The residents from group A averaged a significantly higher score (6.3) than group B (5.0) (P = 0.017). Understanding gait biomechanics is a critical skill for orthopedic surgeons, and residency training often lacks sufficient training in their curriculum. A good understanding of gait analysis allows orthopedic surgeons to analyze gait disturbances and develop patient-specific treatment plans in adult and pediatric populations. This study found improved knowledge of gait patterns amongst orthopedic residents with formal education. Even though the residents with formal education fared better than their counterparts, the overall percent correct was still low (51%). Evidence: Level 3: Prospective Cohort Study.

A Simulation-Based Method for Planning Delivery of TTFields to Brain Tumors

Tumor Treating Fields (TTFields) are alternating electric fields that inhibit tumor growth. TTFields are delivered using two pairs of transducer arrays placed on the skin. Each pair of arrays creates a field oriented along the axis between the arrays. Arrays are placed to generate two fields that are roughly orthogonal to one another. Simulation-based studies have shown that TTFields distribution within the body is heterogeneous, and depends on the location of the arrays, the patient anatomy and the location and size of the tumor. A recent study showed a connection between increased TTFields dose at the tumor bed and extended patient survival in newly diagnosed Glioblastoma (ndGBM) patients [1]. This emphasizes the need for simulation-based algorithms for patient-specific TTFields treatment planning. Here we present a novel algorithm that utilizes numerical simulations in order to maximize TTFields dose in brain tumors. The outline for our optimization algorithm is as follows: First a patient-specific computational head model is created in a semi-automatic manner from patient MRI data. Next, machine vision algorithms are used in order to define an axial slice passing above the eyes and eyebrows. To ensure transducer arrays are placed in a practical manner, only transducer array layouts in which all arrays are placed above this slice are tested. Next, a finite set of pairs of virtual transducer arrays are placed on the head model and the electric field created by each pair simulated using a finite elements method. To ensure maximal delivery of TTFields to the tumor, the arrays are placed such that the “imaginary” line connecting their centers passes through the tumor bed. A set of layouts is created by combining pairs of arrays (all possible combinations of pairs are created), and the dose delivered by the layouts calculated. The layout that maximizes TTFields dose at the tumor bed is found, and a set of small shifts and rotations is applied to the arrays in the chosen layout order eliminate overlaps between the transducers of the different arrays. The algorithm was tested on MRI data from ten Glioblastoma patients. In all cases, the algorithm converged to yield an optimal array layout that could be placed in a practical manner. Generally, the arrays of the optimal layout were located on the scalp in close proximity to the tumor. In all cases, the optimal layouts delivered average TTFields intensities above the therapeutic threshold of 1 V/cm to the tumors. The average dose delivered by the optimal layouts to the tumors were well above 1 mW/cm3, which has been associated with prolonged patient survival. We have developed a novel simulation-based method for optimizing TTFields treatment planning for brain tumors. Incorporation of such algorithms into sophisticated treatment planning systems is expected to improve the management of TTFields patients by the treating physician, ultimately improving patient outcomes.

The Impact of Covariates in Voxel-Wise Lesion-Symptom Mapping.

Background: Voxel-wise lesion-symptom mapping (VLSM) is a statistical technique to infer the structure-function relationship in patients with cerebral strokes. Previous VLSM research suggests that it is important to adjust for various confounders such as lesion size to minimize the inflation of true effects. The aim of this work is to investigate the regional impact of covariates on true effects in VLSM.Methods: A total of 222 follow-up datasets of acute ischemic stroke patients with known NIH Stroke Scale (NIHSS) score at 48-h post-stroke were available for this study. Patient age, lesion volume, and follow-up imaging time were tested for multicollinearity using variance inflation factor analysis and used as covariates in VLSM analyses. Covariate importance maps were computed from the VLSM results by standardizing the beta coefficients of general linear models.Results: Covariates were found to have distinct regional importance with respect to lesion eloquence in the brain. Age has a relatively higher importance in the superior temporal gyrus, inferior parietal lobule, and in the pre- and post-central gyri. Volume explains more variability in the opercular area of the insula, inferior frontal gyrus, and caudate. Follow-up imaging time accounts for most of the variance in the globus pallidus, ventromedial- and dorsolateral putamen, dorsal caudate, pre-motor thalamus, and the dorsal insula.Conclusions: This is the first study investigating and revealing distinctive regional patterns of importance for covariates typically used in VLSM. These covariate importance maps can improve our understanding of the lesion-deficit relationships in patients and could prove valuable for patient-specific treatment and rehabilitation planning.

Quantitative comparison of pre-treatment predictive and post-treatment measured dosimetry for selective internal radiation therapy using cone-beam CT for tumor and liver perfusion territory definition

BackgroundSelective internal radiation therapy (SIRT) is a promising treatment for unresectable hepatic malignancies. Predictive dose calculation based on a simulation using 99mTc-labeled macro-aggregated albumin (99mTc-MAA) before the treatment is considered as a potential tool for patient-specific treatment planning. Post-treatment dose measurement is mainly performed to confirm the planned absorbed dose to the tumor and non-tumor liver volumes. This study compared the predicted and measured absorbed dose distributions.MethodsThirty-one patients (67 tumors) treated by SIRT with resin microspheres were analyzed. Predicted and delivered absorbed dose was calculated using 99mTc-MAA-SPECT and 90Y-TOF-PET imaging. The voxel-level dose distribution was derived using the local deposition model. Liver perfusion territories and tumors have been delineated on contrast-enhanced CBCT images, which have been acquired during the 99mTc-MAA work-up. Several dose-volume histogram (DVH) parameters together with the mean dose for liver perfusion territories and non-tumoral and tumoral compartments were evaluated.ResultsA strong correlation between the predicted and measured mean dose for non-tumoral volume was observed (r = 0.937). The ratio of measured and predicted mean dose to this volume has a first, second, and third interquartile range of 0.83, 1.05, and 1.25. The difference between the measured and predicted mean dose did not exceed 11 Gy. The correlation between predicted and measured mean dose to the tumor was moderate (r = 0.623) with a mean difference of − 9.3 Gy. The ratio of measured and predicted tumor mean dose had a median of 1.01 with the first and third interquartile ranges of 0.58 and 1.59, respectively. Our results suggest that 99mTc-MAA-based dosimetry could predict under or over dosing of the non-tumoral liver parenchyma for almost all cases. For more than two thirds of the tumors, a predictive absorbed dose correctly indicated either good tumor dose coverage or under-dosing of the tumor.ConclusionOur results highlight the predictive value of 99mTc-MAA-based dose estimation to predict non-tumor liver irradiation, which can be applied to prescribe an optimized activity aiming at avoiding liver toxicity. Compared to non-tumoral tissue, a poorer agreement between predicted and measured absorbed dose is observed for tumors.

Rates, predictors, and mortality of sepsis-associated acute kidney injury: a systematic review and meta-analysis

BackgroundDue to the high incidence and mortality of sepsis-associated acute kidney injury, a significant number of studies have explored the causes of sepsis-associated acute kidney injury (AKI). However, the opinions on relevant predictive risk factors remain inconclusive. This study aimed to provide a systematic review and meta-analysis to determine the predisposing factors for sepsis-associated AKI.MethodA systematic literature search was performed in the Medline, Embase, Cochrane Library, PubMed, and Web of Science, databases, with an end-date of 25th May 2019. Valid data were retrieved in compliance with specific inclusion and exclusion criteria.ResultForty-seven observational studies were included for analysis, achieving a cumulative patient number of 55,911. The highest incidence of AKI was caused by septic shock. Thirty-one potential risk factors were included in the meta-analysis. Analysis showed that 20 factors were statistically significant. The odds ratio (OR) and 95% confidence interval (CI), as well as the prevalence of the most frequently-seen predisposing factors for sepsis-associated AKI, were as follows: septic shock [2.88 (2.36–3.52), 60.47%], hypertension [1.43 (1.20–1.70), 38.39%], diabetes mellitus [1.59 (1.47–1.71), 27.57%], abdominal infection [1.44 (1.32–1.58), 30.87%], the administration of vasopressors [2.95 (1.67–5.22), 64.61%], the administration of vasoactive drugs [3.85 (1.89–7.87), 63.22%], mechanical ventilation [1.64 (1.24–2.16), 68.00%], positive results from blood culture [1.60 (1.35–1.89), 41.19%], and a history of smoking [1.60 (1.09–2.36), 43.09%]. Other risk factors included cardiovascular diseases, coronary artery diseases, liver diseases, unknown infections, the administration of diuretics and ACEI/ARB, the infection caused by gram-negative bacteria, and organ transplantation.ConclusionRisk factors of S-AKI arise from a wide range of sources, making it difficult to predict and prevent this condition. Comorbidities, and certain drugs, are the main risk factors for S-AKI. Our review can provide guidance on the application of interventions to reduce the risks associated with sepsis-associated acute kidney injury and can also be used to tailor patient-specific treatment plans and management strategies in clinical practice.

Automatic bi-objective parameter tuning for inverse planning of high-dose-rate prostate brachytherapy

We present an automatic bi-objective parameter-tuning approach for inverse planning methods for high-dose-rate prostate brachytherapy, which aims to overcome the difficult and time-consuming manual parameter tuning that is currently required to obtain patient-specific high-quality treatment plans. We modelled treatment planning as a bi-objective optimization problem, in which dose-volume-based planning criteria related to target coverage are explicitly separated from organ-sparing criteria. When this model is optimized, a large set of high-quality plans with different trade-offs can be obtained. This set can be visualized as an insightful patient-specific trade-off curve. In our parameter-tuning approach, the parameters of inverse planning methods are automatically tuned, aimed to maximize the two objectives of the bi-objective planning model. By generating trade-off curves for different inverse planning methods, their maximally achievable plan quality can be insightfully compared. Automatic parameter tuning furthermore allows to construct standard parameter sets (class solutions) representing different trade-offs in a principled way, which can be directly used in current clinical practice. In this work, we considered the inverse planning methods IPSA and HIPO. Thirty-nine previously treated prostate cancer patients were included. We compared automatic parameter tuning, random parameter sampling, and the maximally achievable plan quality obtained by directly optimizing the bi-objective planning model with the state-of-the-art optimization software GOMEA. We showed that for each patient, a set of plans with a wide range of trade-offs could be obtained using automatic parameter tuning for both IPSA and HIPO. By tuning HIPO, better trade-offs were obtained than by tuning IPSA. For most patients, automatic tuning of HIPO resulted in plans close to the maximally achievable plan quality obtained by optimizing the bi-objective planning model directly. Automatic parameter tuning was shown to improve plan quality significantly compared to random parameter sampling. Finally, from the automatically-tuned plans, three class solutions were successfully constructed representing different trade-offs.

Robustness of Time-Multiplexed Hyperthermia to Temperature Dependent Thermal Tissue Properties

Microwave hyperthermia is a promising cancer treatment used in combination with radio- and chemotherapy. Typically, hyperthermia systems involve several antennas that transfer electromagnetic energy into the tissue. The principal need in hyperthermia treatment is to optimally focus the heating into the target while minimising heating in the surrounding healthy tissue. Patient-specific treatment planning is done to optimize the specific absorption rate and the resulting temperature distribution. Uncertainties associated with the thermal model used for temperature simulations represent an important challenge. Our previous work has demonstrated that the occurrence of hotspots can be reduced and target heating enhanced using time-multiplexed steering procedures. In this paper, the robustness of time-multiplexed hyperthermia against temperature dependent thermal tissue properties is investigated. Temperature simulations are used to predict the time-dependent heating achieved by multiple antenna phase and amplitude configurations that are generated by a multi-objective genetic algorithm and applied sequentially. The proposed strategy is compared with the heating obtained using one single heating setting obtained by particle swarm optimization as typically used in clinical hyperthermia. Thermal performance of the static and time-multiplexed methods are assessed by applying two thermal models, one that uses constant properties of blood perfusion and thermal conductivity of tumor, muscle and fat, and a second one that uses temperature dependent perfusion values. This study shows that time-multiplexed hyperthermia enhances target heating and limits the hotspot appearance regardless of the thermal model used in thermal simulations.

Modelling Curved Contact Flexible Microstrip Applicators for Patient-Specific Superficial Hyperthermia Treatment Planning.

This paper describes a method to reconstruct bendable superficial hyperthermia applicators for routine clinical patient-specific treatment planning. The reconstruction uses a CT scan with a flexible silicone dummy applicator positioned on the patient. The curvature was approximated by two second-degree polynomial functions. A realistic treatment series was mimicked using a standard Alderson radiation therapy phantom and a treatment planning model was reconstructed from a CT scan. The variation among treatment curvatures was compared to the modelled curvature. The mathematical approximation of the applicator curvature was validated for this phantom experiment, as well as for clinical treatments. The average maximum variation among the successive mimicked sessions was 3.67 ± 0.69 mm (range 2.98–4.60 mm). The maximum deviation between the treatment curvature and the modelled curvature was 4.35 mm. Comparing the treatment and approximated curvature yielded a maximum deviation between 2.98 mm and 4.12 mm. For clinical treatments the maximum deviation of the treatment and approximated curvature varied between 0.48 mm and 1.98 mm. These results allow adequate reconstruction of bendable hyperthermia applicators for treatment planning, which can further improve treatment quality, for example by optimizing the water bolus temperature for patient-specific tumor depths. Predictive parameters for hyperthermia treatment outcome can easily be evaluated and compared for various input parameters.

Solid Organ Transplant Is Associated With Increased Morbidity and Mortality in Patients Undergoing One or Two-level Anterior Cervical Decompression and Fusion

Retrospective database review OBJECTIVE.: The aim of this study was to analyze the implications of solid organ transplant (SOT) on postoperative outcomes following elective one or two-level anterior cervical discectomy and fusion (ACDF). Although SOTs have been associated with increased morbidity, postoperative outcomes in SOT recipients undergoing cervical spinal surgery are not well studied. A retrospective database review of Medicare patients younger than 85 years who underwent an elective one to two-level ACDF from 2006-2013 was conducted. Following our exclusion criteria, patients were then divided into the following groups: those with a prior history of kidney, liver, heart or lung transplant (SOT group) and non-SOT patients. Both groups were compared for hospital length of stay, 90-day major medical complications, 90-day hospital readmission, 1-year surgical site infection (SSI), 1-year revision ACDF, and 1-year mortality. A total of 992 (0.5%) SOT recipients (1,144 organs) were identified out of 199,288 ACDF patients. SOT recipients had a significantly longer length of stay (2.32 vs. 5.22 days, p<0.001), higher rate of major medical complications (8.2% vs. 4.5%; OR 1.85, 95% CI 1.45-2.33, p<0.001) and hospital readmission (19.5% vs. 7.5%, OR 2.05, 95% CI 1.74-2.41, p<0.001). In addition, SOT patients had increased mortality within one year of surgery (5.8% vs. 1.3%; OR 3.01, 95% CI 2.26-3.94, p<0.001) compared to non-SOT patients. SOT was not independently associated with SSI (OR 1.25, 95% CI 0.85-1.75, p=0.230), and there was no significant difference in revision rate (0.9% vs. 0.5%; OR 1.54, 95% CI 0.73-2.82, p=0.202) between both groups. SOT is independently associated with longer hospital stay, increased rate of major medical complications, hospital readmission and mortality. Spine surgeons should be aware of the higher rates of morbidity and mortality in these patients and take it into consideration when developing patient-specific treatment plans. 3.

Comparison of treatment plans feasible through AI enabled multidisciplinary online tumor board solution versus NCCN-based clinical decision support system (CDSS).

816 Background: Multidisciplinary tumor boards at Academic Medical Centers (AMC) maximize cancer outcomes. Guidelines based CDSS are alternatives to determine care pathways. Since 2015, 300 AMC cancer experts in USA and India use an AI enabled online tumor board solution, “NAVYA,” to scale low cost access to multidisciplinary expertise, on 1-2 minutes of expert time per decision (ASCO 2017). Methods: GI patients who used NAVYA between 5/1/15-8/31/19 were analyzed. Actionable treatment plans generated by NAVYA were compared to NCCN. Actionable treatment plans include chemotherapy protocols (doses, frequencies), radiation protocols (sites, fractions), etc. Inactionable specialty level decisions (CT-RT vs. surgery) lack specificity. Results: 1302 patients (4638 treatment decisions) were analyzed: 61% (794) male, 80% between age 45 to 75, mostly with Colon, Pancreas, Gallbladder, Rectum, or Stomach cancer; 49.7% non-metastatic. Cohort was comparable to GLOBOCAN estimates. In 82.2% (3812/4638) decisions, NAVYA added value beyond NCCN. First, in 4.5% (212/4638), NAVYA recommended a patient-specific treatment plan that was not part of NCCN. Second, in 3.2% (148/4638), NAVYA recommended treatments plan for clinical scenarios not covered by NCCN, (for eg. 3rd line therapies). Third, in 74.5% (3452/4638), NAVYA used patient specific criteria including resource constraints and patient preference to choose a treatment plan amongst the multiple pathways provided by NCCN and added actionable treatment details. Conclusions: Guideline based CDSS are insufficient to make the vast majority of actionable treatment decisions. Scaling rapid access to multidisciplinary experts is critical. Leapfrogging existing guidelines based CDSS, NAVYA online tumor board makes actionable expert treatment plans possible at a large scale.

Toward Estimating the Uncertainty Associated With Three-Dimensional Geometry Reconstructed From Medical Image Data

Patient-specific computational modeling is increasingly used to assist with visualization, planning, and execution of medical treatments. This trend is placing more reliance on medical imaging to provide accurate representations of anatomical structures. Digital image analysis is used to extract anatomical data for use in clinical assessment/planning. However, the presence of image artifacts, whether due to interactions between the physical object and the scanning modality or the scanning process, can degrade image accuracy. The process of extracting anatomical structures from the medical images introduces additional sources of variability, e.g., when thresholding or when eroding along apparent edges of biological structures. An estimate of the uncertainty associated with extracting anatomical data from medical images would therefore assist with assessing the reliability of patient-specific treatment plans. To this end, two image datasets were developed and analyzed using standard image analysis procedures. The first dataset was developed by performing a "virtual voxelization" of a CAD model of a sphere, representing the idealized scenario of no error in the image acquisition and reconstruction algorithms (i.e., a perfect scan). The second dataset was acquired by scanning three spherical balls using a laboratory-grade CT scanner. For the idealized sphere, the error in sphere diameter was less than or equal to 2% if 5 or more voxels were present across the diameter. The measurement error degraded to approximately 4% for a similar degree of voxelization of the physical phantom. The adaptation of established thresholding procedures to improve segmentation accuracy was also investigated.

Investigation of pelvic symmetry using CAD software.

Severe pelvic fractures often prove difficult for surgeons as they require patient-specific surgical treatment plans and customized equipment. Developing virtual patient-specific 3D pelvis models would ease the surgical planning process and enable development of custom fixation plates. This paper aims to examine pelvic symmetry to conclude whether the contralateral side may be used as a reference model for the fractured side of the pelvis. Fourteen subjects with intact pelvises were involved in this study. CT scans of the pelvises were converted to 3D models and the right sides of the pelvises were reflected and aligned with the left sides. A deviation analysis was then performed for each set of models and results showed that the average root mean square (RMS) of values was 1.14 ± 0.26 mm and the average percentage of points below a deviation threshold of ± 2 mm was 91.9 ± 5.55%. The deviation color maps (DCMs) showed that the largest deviations were on the non-articular surfaces. The volume and surface area of each model were also examined and showed no significant differences between left and right sides. These results indicate that the pelvis displays bilateral symmetry and this concept can be used to develop fully intact patient-specific 3D pelvis models for fracture reconstruction using the unfractured contralateral side. Graphical Abstract.

Emerging Trends of Zygomaticomaxillary Complex Fractures and Their Etiological Analysis in a Tertiary Health Centre from Eastern India: A Retrospective Study.

Maxillofacial region is one of the most complicated arrangements of bony and soft tissues in human body, superiorly bounded by cranial bones and inferiorly by the mandible with interposed dentition. This is one of the reasons which make injuries to this region difficult to manage and often controversial. The severity and patterns of these injuries have changed considerably due to the increasing reliance on fast road transportation facilities and interpersonal violence. The aim of the present study is to analyze retrospectively the changing patterns of mid-face fractures in a sample of patients reporting to Department of Oral and Maxillofacial Surgery in AIIMS Bhubaneswar, India, from April 2017 to April 2018. A performa was prepared which included age, sex, etiology of injury, alcohol influence, and treatment given during hospital stay. Road traffic accidents contributed 68.11% of zygomaticomaxillary complex (ZMC) fractures, followed by assault (18.11%) and fall from height (13.76%). RTA was the major cause of atypical ZMC fractures as compared to conventional ZMC fractures (P<0.001). Due to modern lifestyle and changes in fast transportation, recent fracture patterns do not follow the tell-tale signs or patterns and often need unique and patient-specific treatment plan, making the skills of maxillofacial surgeons the pivotal factor in better prognosis.

A machine learning framework for automated diagnosis and computer-assisted planning in plastic and reconstructive surgery

Current computational tools for planning and simulation in plastic and reconstructive surgery lack sufficient precision and are time-consuming, thus resulting in limited adoption. Although computer-assisted surgical planning systems help to improve clinical outcomes, shorten operation time and reduce cost, they are often too complex and require extensive manual input, which ultimately limits their use in doctor-patient communication and clinical decision making. Here, we present the first large-scale clinical 3D morphable model, a machine-learning-based framework involving supervised learning for diagnostics, risk stratification, and treatment simulation. The model, trained and validated with 4,261 faces of healthy volunteers and orthognathic (jaw) surgery patients, diagnoses patients with 95.5% sensitivity and 95.2% specificity, and simulates surgical outcomes with a mean accuracy of 1.1 ± 0.3 mm. We demonstrate how this model could fully-automatically aid diagnosis and provide patient-specific treatment plans from a 3D scan alone, to help efficient clinical decision making and improve clinical understanding of face shape as a marker for primary and secondary surgery.

P11.37 Evaluating water content and electrical properties at 200 kHz in brain and GBM tumor tissue of three TTFields patients with conventional imaging

Abstract BACKGROUND Electrical properties (EPs) of brain tissue, specifically brain tumors, crucially determine the field distribution of Tumor Treating Fields (TTFields), an anti-mitotic treatment approved for glioblastoma multiforme (GBM). Due to the correlation of TTFields efficacy and field intensity at the tumor region, the knowledge of EPs in each patient is of great importance for patient-specific planning of treatment. Water content electrical properties tomography (wEPT) is a non-invasive imaging technique using water content (WC) maps obtained from rapidly acquired and processed conventional sequences to estimate the EPs of brain tissue at 128 MHz. The WC maps of this approach are constructed from two spin echo sequences similar to a T1 and a PD image. Following previous studies in rat tumor models demonstrating promising wEPT mapping of EPs in the brain at 200, this study examines the feasibility of this approach in human GBM patients. MATERIAL AND METHODS For three patients of the EF-14 trial population, we divided T1 and PD images pixel-by-pixel to obtain the image ratio. Using a transfer function, WC maps were generated and maps of the electrical conductivity σ and the relative permittivity ε r at 200 kHz were calculated with two different equations. RESULTS The median value of estimated WC remains similar in healthy brain tissues among all patients, ~73.5% in the white matter, ~82% in the gray matter. The median values of wEPT-estimated σ at 200 kHz in the white matter is ~0.09 S/m and in the gray matter ~0.18 S/m, corresponding median values of ε r at 200 kHz are ~2100 and ~3000 in white and gray matter respectively. Contrary, in the tumor the spread between the median values of WC and EPs is much higher. Stating the most important findings, in the necrosis median WC are 90.3%, 92.3%, 85.2% in patients 1–3 respectively with corresponding median σ values of 0.494, 0.657, 0.25 S/m. In the enhancing tumor the spread of median WC is even higher (67.2%, 83.6%, 85.5%), yet lower spread but also very heterogeneous median σ values of 0.075 S/m, 0.208, 0.259 S/m are estimated with wEPT. CONCLUSION Our results demonstrate the adaption of wEPT for mapping of WC and EPs at 200 kHz in three human GBM patients. In contrast to the vastly irregular tumor tissue, our estimations in healthy brain tissue are similar between patients and in accordance with EPs experimentally measured during our animal experiments and consistent with reported values in the literature. Hence, wEPT is a promising, fast technique based on regular MRI that might help patient-specific treatment planning of TTFields therapy, although the mapping of tumor tissue needs further confirmation in a greater population and investigations of EPs of excised tumor tissue samples should be conducted.