Publications

2021

Ye Y, Barapatre S, Davis MK, Elliston KO, Davatzikos C, Fedorov A, Fillion-Robin JC, Foster I, Gilbertson JR, Lasso A, Miller J V, Morgan M, Pieper S, Raumann BE, Sarachan BD, Savova G, Silverstein JC, Taylor DP, Zelnis JB, Zhang GQ, Cuticchia J, Becich MJ. Open-source Software Sustainability Models: Initial White Paper From the Informatics Technology for Cancer Research Sustainability and Industry Partnership Working Group. J Med Internet Res. 2021;23(12):e20028.
BACKGROUND: The National Cancer Institute Informatics Technology for Cancer Research (ITCR) program provides a series of funding mechanisms to create an ecosystem of open-source software (OSS) that serves the needs of cancer research. As the ITCR ecosystem substantially grows, it faces the challenge of the long-term sustainability of the software being developed by ITCR grantees. To address this challenge, the ITCR sustainability and industry partnership working group (SIP-WG) was convened in 2019. OBJECTIVE: The charter of the SIP-WG is to investigate options to enhance the long-term sustainability of the OSS being developed by ITCR, in part by developing a collection of business model archetypes that can serve as sustainability plans for ITCR OSS development initiatives. The working group assembled models from the ITCR program, from other studies, and from the engagement of its extensive network of relationships with other organizations (eg, Chan Zuckerberg Initiative, Open Source Initiative, and Software Sustainability Institute) in support of this objective. METHODS: This paper reviews the existing sustainability models and describes 10 OSS use cases disseminated by the SIP-WG and others, including 3D Slicer, Bioconductor, Cytoscape, Globus, i2b2 (Informatics for Integrating Biology and the Bedside) and tranSMART, Insight Toolkit, Linux, Observational Health Data Sciences and Informatics tools, R, and REDCap (Research Electronic Data Capture), in 10 sustainability aspects: governance, documentation, code quality, support, ecosystem collaboration, security, legal, finance, marketing, and dependency hygiene. RESULTS: Information available to the public reveals that all 10 OSS have effective governance, comprehensive documentation, high code quality, reliable dependency hygiene, strong user and developer support, and active marketing. These OSS include a variety of licensing models (eg, general public license version 2, general public license version 3, Berkeley Software Distribution, and Apache 3) and financial models (eg, federal research funding, industry and membership support, and commercial support). However, detailed information on ecosystem collaboration and security is not publicly provided by most OSS. CONCLUSIONS: We recommend 6 essential attributes for research software: alignment with unmet scientific needs, a dedicated development team, a vibrant user community, a feasible licensing model, a sustainable financial model, and effective product management. We also stress important actions to be considered in future ITCR activities that involve the discussion of the sustainability and licensing models for ITCR OSS, the establishment of a central library, the allocation of consulting resources to code quality control, ecosystem collaboration, security, and dependency hygiene.
Schilling KG, Rheault F, Petit L, Hansen CB, Nath V, Yeh FC, Girard G, Barakovic M, Rafael-Patino J, Yu T, Fischi-Gomez E, Pizzolato M, Ocampo-Pineda M, Schiavi S, Canales-Rodríguez EJ, Daducci A, Granziera C, Innocenti G, Thiran JP, Mancini L, Wastling S, Cocozza S, Petracca M, Pontillo G, Mancini M, Vos SB, Vakharia VN, Duncan JS, Melero H, Manzanedo L, Sanz-Morales E, Peña-Melián Á, Calamante F, Attyé A, Cabeen RP, Korobova L, Toga AW, Vijayakumari AA, Parker D, Verma R, Radwan A, Sunaert S, Emsell L, De Luca A, Leemans A, Bajada CJ, Haroon H, Azadbakht H, Chamberland M, Genc S, Tax CMW, Yeh PH, Srikanchana R, Mcknight CD, Yang JYM, Chen J, Kelly CE, Yeh CH, Cochereau J, Maller JJ, Welton T, Almairac F, Seunarine KK, Clark CA, Zhang F, Makris N, Golby A, Rathi Y, Donnell LJO, Xia Y, Aydogan DB, Shi Y, Fernandes FG, Raemaekers M, Warrington S, Michielse S, Ramírez-Manzanares A, Concha L, Aranda R, Meraz MR, Lerma-Usabiaga G, Roitman L, Fekonja LS, Calarco N, Joseph M, Nakua H, Voineskos AN, Karan P, Grenier G, Legarreta JH, Adluru N, Nair VA, Prabhakaran V, Alexander AL, Kamagata K, Saito Y, Uchida W, Andica C, Abe M, Bayrak RG, Wheeler-Kingshott CAMG, Angelo ED, Palesi F, Savini G, Rolandi N, Guevara P, Houenou J, López-López N, Mangin JF, Poupon C, Román C, Vázquez A, Maffei C, Arantes M, Andrade JP, Silva SM, Calhoun VD, Caverzasi E, Sacco S, Lauricella M, Pestilli F, Bullock D, Zhan Y, Brignoni-Perez E, Lebel C, Reynolds JE, Nestrasil I, Labounek R, Lenglet C, Paulson A, Aulicka S, Heilbronner SR, Heuer K, Chandio BQ, Guaje J, Tang W, Garyfallidis E, Raja R, Anderson AW, Landman BA, Descoteaux M. Tractography Dissection Variability: What Happens When 42 Groups Dissect 14 White Matter Bundles on the Same Dataset?. Neuroimage. 2021;243:118502.
White matter bundle segmentation using diffusion MRI fiber tractography has become the method of choice to identify white matter fiber pathways in vivo in human brains. However, like other analyses of complex data, there is considerable variability in segmentation protocols and techniques. This can result in different reconstructions of the same intended white matter pathways, which directly affects tractography results, quantification, and interpretation. In this study, we aim to evaluate and quantify the variability that arises from different protocols for bundle segmentation. Through an open call to users of fiber tractography, including anatomists, clinicians, and algorithm developers, 42 independent teams were given processed sets of human whole-brain streamlines and asked to segment 14 white matter fascicles on six subjects. In total, we received 57 different bundle segmentation protocols, which enabled detailed volume-based and streamline-based analyses of agreement and disagreement among protocols for each fiber pathway. Results show that even when given the exact same sets of underlying streamlines, the variability across protocols for bundle segmentation is greater than all other sources of variability in the virtual dissection process, including variability within protocols and variability across subjects. In order to foster the use of tractography bundle dissection in routine clinical settings, and as a fundamental analytical tool, future endeavors must aim to resolve and reduce this heterogeneity. Although external validation is needed to verify the anatomical accuracy of bundle dissections, reducing heterogeneity is a step towards reproducible research and may be achieved through the use of standard nomenclature and definitions of white matter bundles and well-chosen constraints and decisions in the dissection process.
Novello L, Agarwal N, Vennarini S, Lorentini S, Zacà D, Mussano A, Pasternak O, Jovicich J. Longitudinal Changes in Brain Diffusion MRI Indices during and after Proton Beam Therapy in a Child with Pilocytic Astrocytoma: A Case Report. Diagnostics (Basel). 2021;12(1):26.
Proton beam therapy (PBT) is an effective pediatric brain tumor treatment. However, the resulting microstructural changes within and around irradiated tumors are unknown. We retrospectively applied diffusion tensor imaging (DTI) and free-water imaging (FWI) on diffusion-weighted magnetic resonance imaging (dMRI) data to monitor microstructural changes during the PBT and after 8 months in a pilocytic astrocytoma (PA) and normal-appearing white matter (NAWM). We evaluated the conventional MRI- and dMRI-derived indices from six MRI sessions (t0-t5) in a Caucasian child with a hypothalamic PA: at baseline (t0), during the PBT (t1-t4) and after 8 months (t5). The tumor voxels were classified as "solid" or "fluid" based on the FWI. While the tumor volume remained stable during the PBT, the dMRI analyses identified two different response patterns: (i) an increase in fluid content and diffusivity with anisotropy reductions in the solid voxels at t1, followed by (ii) smaller variations in fluid content but higher anisotropy in the solid voxels at t2-t4. At follow-up (t5), the tumor volume, fluid content, and diffusivity in the solid voxels increased. The NAWM showed dose-dependent microstructural changes. The use of the dMRI and FWI showed complex dynamic microstructural changes in the irradiated mass during the PBT and at follow-up, opening new avenues in our understanding of radiation-induced pathophysiologic mechanisms in tumors and the surrounding tissues.
Herberthson M, Boito D, Haije TD, Feragen A, Westin CF, Özarslan E. Q-Space Trajectory Imaging With Positivity Constraints (QTI+). Neuroimage. 2021;238:118198.
Q-space trajectory imaging (QTI) enables the estimation of useful scalar measures indicative of the local tissue structure. This is accomplished by employing generalized gradient waveforms for diffusion sensitization alongside a diffusion tensor distribution (DTD) model. The first two moments of the underlying DTD are made available by acquisitions at low diffusion sensitivity (b-values). Here, we show that three independent conditions have to be fulfilled by the mean and covariance tensors associated with distributions of symmetric positive semidefinite tensors. We introduce an estimation framework utilizing semi-definite programming (SDP) to guarantee that these conditions are met. Applying the framework on simulated signal profiles for diffusion tensors distributed according to non-central Wishart distributions demonstrates the improved noise resilience of QTI+ over the commonly employed estimation methods. Our findings on a human brain data set also reveal pronounced improvements, especially so for acquisition protocols featuring few number of volumes. Our method’s robustness to noise is expected to not only improve the accuracy of the estimates, but also enable a meaningful interpretation of contrast in the derived scalar maps. The technique’s performance on shorter acquisitions could make it feasible in routine clinical practice.
Hsu TMH, Schawkat K, Berkowitz SJ, Wei JL, Makoyeva A, Legare K, DeCicco C, Paez N, Wu JSH, Szolovits P, Kikinis R, Moser AJ, Goehler A. Artificial Intelligence to Assess Body Composition on Routine Abdominal CT Scans and Predict Mortality in Pancreatic Cancer- a Recipe for Your Local Application. Eur J Radiol. 2021;142:109834.
BACKGROUND: Body composition is associated with mortality; however its routine assessment is too time-consuming. PURPOSE: To demonstrate the value of artificial intelligence (AI) to extract body composition measures from routine studies, we aimed to develop a fully automated AI approach to measure fat and muscles masses, to validate its clinical discriminatory value, and to provide the code, training data and workflow solutions to facilitate its integration into local practice. METHODS: We developed a neural network that quantified the tissue components at the L3 vertebral body level using data from the Liver Tumor Challenge (LiTS) and a pancreatic cancer cohort. We classified sarcopenia using accepted skeletal muscle index cut-offs and visceral fat based its median value. We used Kaplan Meier curves and Cox regression analysis to assess the association between these measures and mortality. RESULTS: Applying the algorithm trained on LiTS data to the local cohort yielded good agreement [>0.8 intraclass correlation (ICC)]; when trained on both datasets, it had excellent agreement (>0.9 ICC). The pancreatic cancer cohort had 136 patients (mean age: 67 ± 11 years; 54% women); 15% had sarcopenia; mean visceral fat was 142 cm2. Concurrent with prior research, we found a significant association between sarcopenia and mortality [mean survival of 15 ± 12 vs. 22 ± 12 (p < 0.05), adjusted HR of 1.58 (95% CI: 1.03-3.33)] but no association between visceral fat and mortality. The detector analysis took 1 ± 0.5 s. CONCLUSIONS: AI body composition analysis can provide meaningful imaging biomarkers from routine exams demonstrating AI’s ability to further enhance the clinical value of radiology reports.
Nilsson M, Eklund G, Szczepankiewicz F, Skorpil M, Bryskhe K, Westin CF, Lindh C, Blomqvist L, Jäderling F. Mapping Prostatic Microscopic Anisotropy Using Linear and Spherical B-Tensor Encoding: A Preliminary Study. Magn Reson Med. 2021;86(4):2025–33.
PURPOSE: Tensor-valued diffusion encoding provides more specific information than conventional diffusion-weighted imaging (DWI), but has mainly been applied in neuroimaging studies. This study aimed to assess its potential for the imaging of prostate cancer (PCa). METHODS: Seventeen patients with histologically proven PCa were enrolled. DWI of the prostate was performed with linear and spherical tensor encoding using a maximal b-value of 1.5 ms/µm2 and a voxel size of 3 × 3 × 4 mm3 . The gamma-distribution model was used to estimate the mean diffusivity (MD), the isotropic kurtosis (MKI ), and the anisotropic kurtosis (MKA ). Regions of interest were placed in MR-defined cancerous tissues, as well as in apparently healthy tissues in the peripheral and transitional zones (PZs and TZs). RESULTS: DWI with linear and spherical encoding yielded different image contrasts at high b-values, which enabled the estimation of MKA and MKI . Compared with healthy tissue (PZs and TZs combined) the cancers displayed a significantly lower MD (P < .05), higher MKI (P < 10-5 ), and lower MKA (P < .05). Compared with the TZ, tissue in the PZ showed lower MD (P < 10-3 ) and higher MKA (P < 10-3 ). No significant differences were found between cancers of different Gleason scores, possibly because of the limited sample size. CONCLUSION: Tensor-valued diffusion encoding enabled mapping of MKA and MKI in the prostate. The elevated MKI in PCa compared with normal tissues suggests an elevated heterogeneity in the cancers. Increased in-plane resolution could improve tumor delineation in future studies.
Martins JP de A, Nilsson M, Lampinen B, Palombo M, While PT, Westin CF, Szczepankiewicz F. Neural Networks for Parameter Estimation in Microstructural MRI: Application to a Diffusion-Relaxation Model of White Matter. Neuroimage. 2021;244:118601.
Specific features of white matter microstructure can be investigated by using biophysical models to interpret relaxation-diffusion MRI brain data. Although more intricate models have the potential to reveal more details of the tissue, they also incur time-consuming parameter estimation that may converge to inaccurate solutions due to a prevalence of local minima in a degenerate fitting landscape. Machine-learning fitting algorithms have been proposed to accelerate the parameter estimation and increase the robustness of the attained estimates. So far, learning-based fitting approaches have been restricted to microstructural models with a reduced number of independent model parameters where dense sets of training data are easy to generate. Moreover, the degree to which machine learning can alleviate the degeneracy problem is poorly understood. For conventional least-squares solvers, it has been shown that degeneracy can be avoided by acquisition with optimized relaxation-diffusion-correlation protocols that include tensor-valued diffusion encoding. Whether machine-learning techniques can offset these acquisition requirements remains to be tested. In this work, we employ artificial neural networks to vastly accelerate the parameter estimation for a recently introduced relaxation-diffusion model of white matter microstructure. We also develop strategies for assessing the accuracy and sensitivity of function fitting networks and use those strategies to explore the impact of the acquisition protocol. The developed learning-based fitting pipelines were tested on relaxation-diffusion data acquired with optimal and sub-optimal acquisition protocols. Networks trained with an optimized protocol were observed to provide accurate parameter estimates within short computational times. Comparing neural networks and least-squares solvers, we found the performance of the former to be less affected by sub-optimal protocols; however, model fitting networks were still susceptible to degeneracy issues and their use could not fully replace a careful design of the acquisition protocol.
Levitt JJ, Zhang F, Vangel M, Nestor PG, Rathi Y, Kubicki M, Shenton ME, Donnell LJO. The Organization of Frontostriatal Brain Wiring in Healthy Subjects Using a Novel Diffusion Imaging Fiber Cluster Analysis. Cereb Cortex. 2021;31(12):5308–18.
To assess normal organization of frontostriatal brain wiring, we analyzed diffusion magnetic resonance imaging (dMRI) scans in 100 young adult healthy subjects (HSs). We identified fiber clusters intersecting the frontal cortex and caudate, a core component of associative striatum, and quantified their degree of deviation from a strictly topographic pattern. Using whole brain dMRI tractography and an automated tract parcellation clustering method, we extracted 17 white matter fiber clusters per hemisphere connecting the frontal cortex and caudate. In a novel approach to quantify the geometric relationship among clusters, we measured intercluster endpoint distances between corresponding cluster pairs in the frontal cortex and caudate. We show first, the overall frontal cortex wiring pattern of the caudate deviates from a strictly topographic organization due to significantly greater convergence in regionally specific clusters; second, these significantly convergent clusters originate in subregions of ventrolateral, dorsolateral, and orbitofrontal prefrontal cortex (PFC); and, third, a similar organization in both hemispheres. Using a novel tractography method, we find PFC-caudate brain wiring in HSs deviates from a strictly topographic organization due to a regionally specific pattern of cluster convergence. We conjecture cortical subregions projecting to the caudate with greater convergence subserve functions that benefit from greater circuit integration.
Diffusion MRI (dMRI) can probe the tissue microstructure but suffers from low signal-to-noise ratio (SNR) whenever high resolution is combined with high diffusion encoding strengths. Low SNR leads to poor precision as well as poor accuracy of the diffusion-weighted signal; the latter is caused by the rectified noise floor and can be observed as a positive bias in magnitude signal. Super-resolution techniques may facilitate a beneficial tradeoff between bias and resolution by allowing acquisition at low spatial resolution and high SNR, whereafter high spatial resolution is recovered by image reconstruction. In this work, we describe a super-resolution reconstruction framework for dMRI and investigate its performance with respect to signal accuracy and precision. Using phantom experiments and numerical simulations, we show that the super-resolution approach improves accuracy by facilitating a more beneficial trade-off between spatial resolution and diffusion encoding strength before the noise floor affects the signal. By contrast, precision is shown to have a less straightforward dependency on acquisition, reconstruction, and intrinsic tissue parameters. Indeed, we find a gain in precision from super-resolution reconstruction is substantial only when some spatial resolution is sacrificed. Finally, we deployed super-resolution reconstruction in a healthy brain for the challenging combination of spherical b-tensor encoding at ultra-high b-values and high spatial resolution-a configuration that produces a unique contrast that emphasizes tissue in which diffusion is restricted in all directions. This demonstration showcased that super-resolution reconstruction enables a vastly superior image contrast compared to conventional imaging, facilitating investigations that would otherwise have prohibitively low SNR, resolution or require non-conventional MRI hardware.