In JMIR medical education

BACKGROUND : Prospective physicians are expected to find artificial intelligence (AI) to be a key technology in their future practice. This transformative change has caught the attention of scientists, educators, and policy makers alike, with substantive efforts dedicated to the selection and delivery of AI topics and competencies in the medical curriculum. Less is known about the behavioral perspective or the necessary and sufficient preconditions for medical students' intention to use AI in the first place.

OBJECTIVE : Our study focused on medical students' knowledge, experience, attitude, and beliefs related to AI and aimed to understand whether they are necessary conditions and form sufficient configurations of conditions associated with behavioral intentions to use AI in their future medical practice.

METHODS : We administered a 2-staged questionnaire operationalizing the variables of interest (ie, knowledge, experience, attitude, and beliefs related to AI, as well as intention to use AI) and recorded 184 responses at t₀ (February 2020, before the COVID-19 pandemic) and 138 responses at t₁ (January 2021, during the COVID-19 pandemic). Following established guidelines, we applied necessary condition analysis and fuzzy-set qualitative comparative analysis to analyze the data.

RESULTS : Findings from the fuzzy-set qualitative comparative analysis show that the intention to use AI is only observed when students have a strong belief in the role of AI (individually necessary condition); certain AI profiles, that is, combinations of knowledge and experience, attitudes and beliefs, and academic level and gender, are always associated with high intentions to use AI (equifinal and sufficient configurations); and profiles associated with nonhigh intentions cannot be inferred from profiles associated with high intentions (causal asymmetry).

CONCLUSIONS : Our work contributes to prior knowledge by showing that a strong belief in the role of AI in the future of medical professions is a necessary condition for behavioral intentions to use AI. Moreover, we suggest that the preparation of medical students should go beyond teaching AI competencies and that educators need to account for the different AI profiles associated with high or nonhigh intentions to adopt AI.

Wagner Gerit, Raymond Louis, Paré Guy

2023-Mar-22

artificial intelligence, attitudes and beliefs, behavioral intentions, fsQCA, fuzzy-set qualitative comparative analysis, knowledge and experience, medical education

In JMIR public health and surveillance

BACKGROUND : Skeletal muscle and BMI are essential prognostic factors for survival in colorectal cancer (CRC). However, there is a lack of understanding due to scarce studies on the continuous aspects of these variables.

OBJECTIVE : This study aimed to evaluate the prognostic impact of the initial status and trajectories of muscle and BMI on overall survival (OS) and assess whether these 4 profiles within 1 year can represent the profiles 6 years later.

METHODS : We analyzed 4056 newly diagnosed patients with CRC between 2010 to 2020. The volume of the muscle with 5-mm thickness at the third lumbar spine level was measured using a pretrained deep learning algorithm. The skeletal muscle volume index (SMVI) was defined as the muscle volume divided by the square of the height. The correlation between BMI status at the first, third, and sixth years of diagnosis was analyzed and assessed similarly for muscle profiles. Prognostic significances of baseline BMI and SMVI and their 1-year trajectories for OS were evaluated by restricted cubic spline analysis and survival analysis. Patients were categorized based on these 4 dimensions, and prognostic risks were predicted and demonstrated using heat maps.

RESULTS : Trajectories of SMVI were categorized as decreased (812/4056, 20%), steady (2014/4056, 49.7%), or increased (1230/4056, 30.3%). Similarly, BMI trajectories were categorized as decreased (792/4056, 19.5%), steady (2253/4056, 55.5%), or increased (1011/4056, 24.9%). BMI and SMVI values in the first year after diagnosis showed a statistically significant correlation with those in the third and sixth years (P<.001). Restricted cubic spline analysis showed a nonlinear relationship between baseline BMI and SMVI change ratio and OS; BMI, in particular, showed a U-shaped correlation. According to survival analysis, increased BMI (hazard ratio [HR] 0.83; P=.02), high baseline SMVI (HR 0.82; P=.04), and obesity stage 1 (HR 0.80; P=.02) showed a favorable impact, whereas decreased SMVI trajectory (HR 1.31; P=.001), decreased BMI (HR 1.23; P=.02), and initial underweight (HR 1.38; P=.02) or obesity stages 2-3 (HR 1.79; P=.01) were negative prognostic factors for OS. Considered simultaneously, BMI >30 kg/m² with a low SMVI at the time of diagnosis resulted in the highest mortality risk. We observed improved survival in patients with increased muscle mass without BMI loss compared to those with steady muscle mass and BMI.

CONCLUSIONS : Profiles within 1 year of both BMI and muscle were surrogate indicators for predicting the later profiles. Continuous trajectories of body and muscle mass are independent prognostic factors of patients with CRC. An automatic algorithm provides a unique opportunity to conduct longitudinal evaluations of body compositions. Further studies to understand the complicated natural courses of muscularity and adiposity are necessary for clinical application.

Seo Dongjin, Kim Han Sang, Ahn Joong Bae, Park Yu Rang

2023-Mar-22

BMI, SMVI, body mass index, colorectal cancer, deep neural network model, skeletal muscle, skeletal muscle volume index

In Translational vision science & technology

PURPOSE : To develop and validate a fully automated program for choroidal structure analysis within a 1500-µm-wide region of interest centered on the fovea (deep learning-based choroidal structure assessment program [DCAP]).

METHODS : A total of 2162 fovea-centered radial swept-source optical coherence tomography (SS-OCT) B-scans from 162 myopic children with cycloplegic spherical equivalent refraction ranging from -1.00 to -5.00 diopters were collected to develop the DCAP. Medical Transformer network and Small Attention U-Net were used to automatically segment the choroid boundaries and the nulla (the deepest point within the fovea). Automatic denoising based on choroidal vessel luminance and binarization were applied to isolate choroidal luminal/stromal areas. To further compare the DCAP with the traditional handcrafted method, the luminal/stromal areas and choroidal vascularity index (CVI) values for 20 OCT images were measured by three graders and the DCAP separately. Intraclass correlation coefficients (ICCs) and limits of agreement were used for agreement analysis.

RESULTS : The mean ± SD pixel-wise distances from the predicted choroidal inner, outer boundary, and nulla to the ground truth were 1.40 ± 1.23, 5.40 ± 2.24, and 1.92 ± 1.13 pixels, respectively. The mean times required for choroidal structure analysis were 1.00, 438.00 ± 75.88, 393.25 ± 78.77, and 410.10 ± 56.03 seconds per image for the DCAP and three graders, respectively. Agreement between the automatic and manual area measurements was excellent (ICCs > 0.900) but poor for the CVI (0.627; 95% confidence interval, 0.279-0.832). Additionally, the DCAP demonstrated better intersession repeatability.

CONCLUSIONS : The DCAP is faster than manual methods. Also, it was able to reduce the intra-/intergrader and intersession variations to a small extent.

TRANSLATIONAL RELEVANCE : The DCAP could aid in choroidal structure assessment.

Xuan Meng, Wang Wei, Shi Danli, Tong James, Zhu Zhuoting, Jiang Yu, Ge Zongyuan, Zhang Jian, Bulloch Gabriella, Peng Guankai, Meng Wei, Li Cong, Xiong Ruilin, Yuan Yixiong, He Mingguang

2023-Mar-01

In Translational vision science & technology

PURPOSE : The purpose of this study was to build a deep-learning model that automatically analyzes cataract surgical videos for the locations of surgical landmarks, and to derive skill-related motion metrics.

METHODS : The locations of the pupil, limbus, and 8 classes of surgical instruments were identified by a 2-step algorithm: (1) mask segmentation and (2) landmark identification from the masks. To perform mask segmentation, we trained the YOLACT model on 1156 frames sampled from 268 videos and the public Cataract Dataset for Image Segmentation (CaDIS) dataset. Landmark identification was performed by fitting ellipses or lines to the contours of the masks and deriving locations of interest, including surgical tooltips and the pupil center. Landmark identification was evaluated by the distance between the predicted and true positions in 5853 frames of 10 phacoemulsification video clips. We derived the total path length, maximal speed, and covered area using the tip positions and examined the correlation with human-rated surgical performance.

RESULTS : The mean average precision score and intersection-over-union for mask detection were 0.78 and 0.82. The average distance between the predicted and true positions of the pupil center, phaco tip, and second instrument tip was 5.8, 9.1, and 17.1 pixels. The total path length and covered areas of these landmarks were negatively correlated with surgical performance.

CONCLUSIONS : We developed a deep-learning method to localize key anatomical portions of the eye and cataract surgical tools, which can be used to automatically derive metrics correlated with surgical skill.

TRANSLATIONAL RELEVANCE : Our system could form the basis of an automated feedback system that helps cataract surgeons evaluate their performance.

Yeh Hsu-Hang, Jain Anjal M, Fox Olivia, Sebov Kostya, Wang Sophia Y

2023-Mar-01

In JAMA network open

IMPORTANCE : Patients with newly diagnosed locally advanced cervical carcinomas or recurrences after surgery undergoing radiochemotherapy whose tumor is unsuited for a brachytherapy boost need high-dose percutaneous radiotherapy with small margins to compensate for clinical target volume deformations and set-up errors. Cone-beam computed tomography-based online adaptive radiotherapy (ART) has the potential to reduce planning target volume (PTV) margins below 5 mm for these tumors.

OBJECTIVE : To compare online ART technologies with image-guided radiotherapy (IGRT) for gynecologic tumors.

DESIGN, SETTING, AND PARTICIPANTS : This comparative effectiveness study comprised all 7 consecutive patients with gynecologic tumors who were treated with ART with artificial intelligence segmentation from January to May 2022 at the West German Cancer Center. All adapted treatment plans were reviewed for the new scenario of organs at risk and target volume. Dose distributions of adapted and scheduled plans optimized on the initial planning computed tomography scan were compared.

EXPOSURE : Online ART for gynecologic tumors.

MAIN OUTCOMES AND MEASURES : Target dose coverage with ART compared with IGRT for PTV margins of 5 mm or less in terms of the generalized equivalent uniform dose (gEUD) without increasing the gEUD for the organs at risk (bladder and rectum).

RESULTS : The first 10 treatment series among 7 patients (mean [SD] age, 65.7 [16.5] years) with gynecologic tumors from a prospective observational trial performed with ART were compared with IGRT. For a clinical PTV margin of 5 mm, IGRT was associated with a median gEUD decrease in the interfractional clinical target volume of -1.5% (90% CI, -31.8% to 2.9%) for all fractions in comparison with the planned dose distribution. Online ART was associated with a decrease of -0.02% (90% CI, -3.2% to 1.5%), which was less than the decrease with IGRT (P < .001). This was not associated with an increase in the gEUD for the bladder or rectum. For a PTV margin of 0 mm, the median gEUD deviation with IGRT was -13.1% (90% CI, -47.9% to 1.6%) compared with 0.1% (90% CI, -2.3% to 6.6%) with ART (P < .001). The benefit associated with ART was larger for a PTV margin of 0 mm than of 5 mm (P = .004) due to spreading of the cold spot at the clinical target volume margin from fraction to fraction with a median SD of 2.4 cm (90% CI, 1.9-3.4 cm) for all patients.

CONCLUSIONS AND RELEVANCE : This study suggests that ART is associated with an improvement in the percentage deviation of gEUD for the interfractional clinical target volume compared with IGRT. As the gain of ART depends on fractionation and PTV margin, a strategy is proposed here to switch from IGRT to ART, if the delivered gEUD distribution becomes unfavorable in comparison with the expected distribution during the course of treatment.

Guberina Maja, Santiago Garcia Alina, Khouya Aymane, Pöttgen Christoph, Holubyev Kostyantyn, Ringbaek Toke Printz, Lachmuth Manfred, Alberti Yasemin, Hoffmann Christian, Hlouschek Julian, Gauler Thomas, Lübcke Wolfgang, Indenkämpen Frank, Stuschke Martin, Guberina Nika

2023-Mar-01

In JAMA network open

IMPORTANCE : Older adults with advanced cancer who have high pretreatment symptom severity often experience adverse events during cancer treatments. Unsupervised machine learning may help stratify patients into different risk groups.

OBJECTIVE : To evaluate whether clusters identified from baseline patient-reported symptom severity were associated with adverse outcomes.

DESIGN, SETTING, AND PARTICIPANTS : This secondary analysis of the Geriatric Assessment Intervention for Reducing Toxicity in Older Patients With Advanced Cancer (GAP70+) Trial (2014-2019) included patients who completed the National Cancer Institute Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE) before starting a new cancer treatment regimen and received care at community oncology sites across the United States. An unsupervised machine learning algorithm (k-means with Euclidean distance) clustered patients based on similarities of baseline symptom severities. Clustering variables included severity items of 24 PRO-CTCAE symptoms (range, 0-4; corresponding to none, mild, moderate, severe, and very severe). Total severity score was calculated as the sum of 24 items (range, 0-96). Whether the clusters were associated with unplanned hospitalization, death, and toxic effects was then examined. Analyses were conducted in January and February 2022.

EXPOSURES : Symptom severity.

MAIN OUTCOMES AND MEASURES : Unplanned hospitalization over 3 months (primary), all-cause mortality over 1 year, and any clinician-rated grade 3 to 5 toxic effect over 3 months.

RESULTS : Of 718 enrolled patients, 706 completed baseline PRO-CTCAE and were included (mean [SD] age, 77.2 [5.5] years, 401 [56.8%] male patients; 51 [7.2%] Black and 619 [87.8%] non-Hispanic White patients; 245 [34.7%] with gastrointestinal cancer; 175 [24.8%] with lung cancer; mean [SD] impaired Geriatric Assessment domains, 4.5 [1.6]). The algorithm classified 310 (43.9%), 295 (41.8%), and 101 (14.3%) into low-, medium-, and high-severity clusters (within-cluster mean [SD] severity scores: low, 6.3 [3.4]; moderate, 16.6 [4.3]; high, 29.8 [7.8]; P < .001). Controlling for sociodemographic variables, clinical factors, study group, and practice site, compared with patients in the low-severity cluster, those in the moderate-severity cluster were more likely to experience hospitalization (risk ratio, 1.36; 95% CI, 1.01-1.84; P = .046). Moderate- and high-severity clusters were associated with a higher risk of death (moderate: hazard ratio, 1.31; 95% CI, 1.01-1.69; P = .04; high: hazard ratio, 2.00; 95% CI, 1.43-2.78; P < .001), but not toxic effects.

CONCLUSIONS AND RELEVANCE : In this study, unsupervised machine learning partitioned patients into distinct symptom severity clusters; patients with higher pretreatment severity were more likely to experience hospitalization and death.

TRIAL REGISTRATION : ClinicalTrials.gov Identifier: NCT02054741.

Xu Huiwen, Mohamed Mostafa, Flannery Marie, Peppone Luke, Ramsdale Erika, Loh Kah Poh, Wells Megan, Jamieson Leah, Vogel Victor G, Hall Bianca Alexandra, Mustian Karen, Mohile Supriya, Culakova Eva

2023-Mar-01