Predicting lymph node metastasis from primary tumor histology and clinicopathologic factors in colorectal cancer using deep learning.

BACKGROUND: Presence of lymph node metastasis (LNM) influences prognosis and clinical decision-making in colorectal cancer. However, detection of LNM is variable and depends on a number of external factors. Deep learning has shown success in computational pathology, but has struggled to boost performance when combined with known predictors. METHODS: Machine-learned features are created by clustering deep learning embeddings of small patches of tumor in colorectal cancer via k-means, and then selecting the top clusters that add predictive value to a logistic regression model when combined with known baseline clinicopathological variables. We then analyze performance of logistic regression models trained with and without these machine-learned features in combination with the baseline variables. RESULTS: The machine-learned extracted features provide independent signal for the presence of LNM (AUROC: 0.638, 95% CI: [0.590, 0.683]). Furthermore, the machine-learned features add predictive value to the set of 6 clinicopathologic variables in an external validation set (likelihood ratio test, p < 0.00032; AUROC: 0.740, 95% CI: [0.701, 0.780]). A model incorporating these features can also further risk-stratify patients with and without identified metastasis (p < 0.001 for both stage II and stage III). CONCLUSION: This work demonstrates an effective approach to combine deep learning with established clinicopathologic factors in order to identify independently informative features associated with LNM. Further work building on these specific results may have important impact in prognostication and therapeutic decision making for LNM. Additionally, this general computational approach may prove useful in other contexts.

When colorectal cancers spread to the lymph nodes, it can indicate a poorer prognosis. However, detecting lymph node metastasis (spread) can be difficult and depends on a number of factors such as how samples are taken and processed. Here, we show that machine learning, which involves computer software learning from patterns in data, can predict lymph node metastasis in patients with colorectal cancer from the microscopic appearance of their primary tumor and the clinical characteristics of the patients. We also show that the same approach can predict patient survival. With further work, our approach may help clinicians to inform patients about their prognosis and decide on appropriate treatments.

Deep Learning and Imaging for the Orthopaedic Surgeon: How Machines "Read" Radiographs.

➤: In the not-so-distant future, orthopaedic surgeons will be exposed to machines that begin to automatically "read" medical imaging studies using a technology called deep learning. ➤: Deep learning has demonstrated remarkable progress in the analysis of medical imaging across a range of modalities that are commonly used in orthopaedics, including radiographs, computed tomographic scans, and magnetic resonance imaging scans. ➤: There is a growing body of evidence showing clinical utility for deep learning in musculoskeletal radiography, as evidenced by studies that use deep learning to achieve an expert or near-expert level of performance for the identification and localization of fractures on radiographs. ➤: Deep learning is currently in the very early stages of entering the clinical setting, involving validation and proof-of-concept studies for automated medical image interpretation. ➤: The success of deep learning in the analysis of medical imaging has been propelling the field forward so rapidly that now is the time for surgeons to pause and understand how this technology works at a conceptual level, before (not after) the technology ends up in front of us and our patients. That is the purpose of this article.

Automatic Hip Fracture Identification and Functional Subclassification with Deep Learning.

PURPOSE: To investigate the feasibility of automatic identification and classification of hip fractures using deep learning, which may improve outcomes by reducing diagnostic errors and decreasing time to operation. MATERIALS AND METHODS: Hip and pelvic radiographs from 1118 studies were reviewed, and 3026 hips were labeled via bounding boxes and classified as normal, displaced femoral neck fracture, nondisplaced femoral neck fracture, intertrochanteric fracture, previous open reduction and internal fixation, or previous arthroplasty. A deep learning-based object detection model was trained to automate the placement of the bounding boxes. A Densely Connected Convolutional Neural Network (or DenseNet) was trained on a subset of the bounding box images, and its performance was evaluated on a held-out test set and by comparison on a 100-image subset with two groups of human observers: fellowship-trained radiologists and orthopedists; senior residents in emergency medicine, radiology, and orthopedics. RESULTS: The binary accuracy for detecting a fracture of this model was 93.7% (95% confidence interval [CI]: 90.8%, 96.5%), with a sensitivity of 93.2% (95% CI: 88.9%, 97.1%) and a specificity of 94.2% (95% CI: 89.7%, 98.4%). Multiclass classification accuracy was 90.8% (95% CI: 87.5%, 94.2%). When compared with the accuracy of human observers, the accuracy of the model achieved an expert-level classification, at the very least, under all conditions. Additionally, when the model was used as an aid, human performance improved, with aided resident performance approximating unaided fellowship-trained expert performance in the multiclass classification. CONCLUSION: A deep learning model identified and classified hip fractures with expert-level performance, at the very least, and when used as an aid, improved human performance, with aided resident performance approximating that of unaided fellowship-trained attending physicians.Supplemental material is available for this article.© RSNA, 2020.

Hip Fractures and the Bundle: A Cost Analysis of Patients Undergoing Hip Arthroplasty for Femoral Neck Fracture vs Degenerative Joint Disease.

BACKGROUND: The purpose of this study is to determine whether episode Target Prices in the Bundled Payment for Care Improvement (BPCI) initiative sufficiently match the complexities and expenses expected for patients undergoing hip arthroplasty for femoral neck fracture (FNF) as compared to hip degenerative joint disease (DJD). METHODS: Claims data under BPCI Model 2 were collected for patients undergoing hip arthroplasty at a single institution over a 2-year period. Payments from the index hospitalization to 90 days postoperatively were aggregated by Medicare Severity Diagnosis-Related Group (469 or 470), indication (DJD vs FNF), and categorized as index procedure, postacute services, and related hospital readmissions. Actual episode costs and Target Prices were compared in both the FNF and DJD cohorts undergoing hip arthroplasty to gauge the cost discrepancy in each group. RESULTS: A total of 183 patients were analyzed (31 with FNFs, 152 with DJD). In total, the FNF cohort incurred a $415,950 loss under the current episode Target Prices, whereas the DJD cohort incurred a $172,448 gain. Episode Target Prices were significantly higher than actual episode prices for the DJD cohort ($32,573 vs $24,776, P < .001). However, Target Prices were significantly lower than actual episode prices for the FNF cohort ($32,672 vs $49,755, P = .021). CONCLUSION: Episode Target Prices in the current BPCI model fall dramatically short of the actual expenses incurred by FNF patients undergoing hip arthroplasty. Better risk-adjusting Target Prices for this fragile population should be considered to avoid disincentives and delays in care.

Predictors of surgical revision after in situ decompression of the ulnar nerve.

BACKGROUND: This study was performed to identify factors associated with the need for revision surgery after in situ decompression of the ulnar nerve for cubital tunnel syndrome. METHODS: This case-control investigation examined all patients treated at one institution with open in situ decompression for cubital tunnel syndrome between 2006 and 2011. The case patients were 44 failed decompressions that required revision, and the controls were 79 randomly selected patients treated with a single operation. Demographic data and disease-specific data were extracted from the medical records. The rate of revision surgery after in situ decompression was determined from our 5-year experience. A multivariate logistic regression model was used based on univariate testing to determine predictors of revision cubital tunnel surgery. RESULTS: Revision surgery was required in 19% (44 of 231) of all in situ decompressions performed during the study period. Predictors of revision surgery included a history of elbow fracture or dislocation (odds ratio [OR], 7.1) and McGowan stage I disease (OR, 3.2). Concurrent surgery with in situ decompression was protective against revision surgery (OR, 0.19). DISCUSSION: The rate of revision cubital tunnel surgery after in situ nerve decompression should be weighed against the benefits of a less invasive procedure compared with transposition. When considering in situ ulnar nerve decompression, prior elbow fracture as well as patients requesting surgery for mild clinically graded disease should be viewed as risk factors for revision surgery. Patient factors often considered relevant to surgical outcomes, including age, sex, body mass index, tobacco use, and diabetes status, were not associated with a greater likelihood of revision cubital tunnel surgery.

Outcomes of revision surgery for cubital tunnel syndrome.

PURPOSE: To compare both validated patient-rated and objective outcomes of patients following revision cubital tunnel surgery to a similar group of patients who underwent primary surgery. METHODS: This case-control investigation enrolled 56 patients treated surgically for cubital tunnel syndrome (28 revision cases, 28 primary controls) at a tertiary center. Patients with a minimum of 2 years of follow-up were eligible. All patients completed an in-office study evaluation. Revision participants represented 55% of potential patients in our practice and controls (treated only with primary surgery) were chosen at random from our practice to reach a 1:1 case to control ratio. Preoperative McGowan grading was confirmed similar between the groups. Outcome measures included validated patient outcome questionnaires (Patient-Rated Elbow Evaluation, Levine-Katz questionnaire), symptoms, and physical examination findings. Statistical analyses were conducted to compare the patient groups. RESULTS: Despite 79% of revision patients reporting symptomatic improvement, revision patients reported worse outcomes on all measured standardized questionnaires compared with primary patients. The Levine-Katz questionnaire indicated mild residual symptoms in the primary group (1.6) versus moderate remaining symptoms following revision surgery (2.3). The Patient-Rated Elbow Evaluation also indicated superior results for the control group (9 ± 10) compared with the revision group (32 ± 22). Revision patients had a higher frequency of constant symptoms, elevated 2-point discrimination, and diminished pinch strength. McGowan grading improved after 25% of revision surgeries versus 64% of primary surgeries, and 21% of revision patients had deterioration of their McGowan grade. CONCLUSIONS: Subjective and objective outcomes of revision patients in this cohort were inferior to outcomes of similar patients following primary surgery. Revision surgery can be offered in the setting of persistent or recurrent symptoms that are unexplained by an alternative diagnosis, but patients should be counseled that complete resolution of symptoms is unlikely. TYPE OF STUDY/LEVEL OF EVIDENCE: Therapeutic III.