Research Spotlight

Posted March 16th 2021

A contextual detector of surgical tools in laparoscopic videos using deep learning.

Ganesh Sankaranarayanan Ph.D.

Ganesh Sankaranarayanan Ph.D.

Namazi, B., Sankaranarayanan, G. and Devarajan, V. (2021). “A contextual detector of surgical tools in laparoscopic videos using deep learning.” Surg Endosc Feb 8. [Epub ahead of print].

Full text of this article.

BACKGROUND: The complexity of laparoscopy requires special training and assessment. Analyzing the streaming videos during the surgery can potentially improve surgical education. The tedium and cost of such an analysis can be dramatically reduced using an automated tool detection system, among other things. We propose a new multilabel classifier, called LapTool-Net to detect the presence of surgical tools in each frame of a laparoscopic video. METHODS: The novelty of LapTool-Net is the exploitation of the correlations among the usage of different tools and, the tools and tasks-i.e., the context of the tools’ usage. Towards this goal, the pattern in the co-occurrence of the tools is utilized for designing a decision policy for the multilabel classifier based on a Recurrent Convolutional Neural Network (RCNN), which is trained in an end-to-end manner. In the post-processing step, the predictions are corrected by modeling the long-term tasks’ order with an RNN. RESULTS: LapTool-Net was trained using publicly available datasets of laparoscopic cholecystectomy, viz., M2CAI16 and Cholec80. For M2CAI16, our exact match accuracies (when all the tools in one frame are predicted correctly) in online and offline modes were 80.95% and 81.84% with per-class F1-score of 88.29% and 90.53%. For Cholec80, the accuracies were 85.77% and 91.92% with F1-scores if 93.10% and 96.11% for online and offline, respectively. CONCLUSIONS: The results show LapTool-Net outperformed state-of-the-art methods significantly, even while using fewer training samples and a shallower architecture. Our context-aware model does not require expert’s domain-specific knowledge, and the simple architecture can potentially improve all existing methods.


Posted March 16th 2021

Huge Right Ventricular Outflow Tract Aneurysm Late Following Total Repair of Tetralogy of Fallot Leading to Orthotopic Heart Transplantation.

William C. Roberts M.D.

William C. Roberts M.D.

Roberts, W.C. and Siddiqui, S. (2021). “Huge Right Ventricular Outflow Tract Aneurysm Late Following Total Repair of Tetralogy of Fallot Leading to Orthotopic Heart Transplantation.” Cardiovasc Pathol Mar 2;107332. [Epub ahead of print].

Full text of this article.

Described herein are 3 patients who underwent successful orthotopic heart transplantation (OHT) because of huge, calcified right ventricular outflow tract (RVOT) aneurysms after repair of tetralogy of Fallot 35, 43, and 59 years earlier. Two of the 3 patients developed recurring episodes of ventricular tachycardia simulating arrhythmogenic right ventricular cardiomyopathy.


Posted March 16th 2021

A Multicenter Investigation of the Hemodynamic Effects of Induction Agents for Trauma Rapid Sequence Intubation.

Justin Regner M.D.

Justin Regner M.D.

Leede, E., Kempema, J., Wilson, C., Rios Tovar, A.J., Cook, A., Fox, E., Regner, J., Richmond, R., Carrick, M. and Brown, C.V.R. (2021). “A Multicenter Investigation of the Hemodynamic Effects of Induction Agents for Trauma Rapid Sequence Intubation.” J Trauma Acute Care Surg Mar 2. [Epub ahead of print].

Full text of this article.

BACKGROUND: Several options exist for induction agents during rapid sequence intubation (RSI) in trauma patients, including etomidate, ketamine, and propofol. These drugs have reported variable hemodynamic effects (hypotension with propofol and sympathomimetic effects with ketamine) that could affect trauma resuscitations. The purpose of this study was to compare the hemodynamic effects of these three induction agents during emergency department RSI in adult trauma. We hypothesized that these drugs would display a differing hemodynamic profile during RSI. METHODS: We performed a retrospective (2014-2019), multicenter trial of adult (>/=18 years) trauma patients admitted to eight ACS-verified Level 1 trauma centers who underwent emergency department RSI. Variables collected included systolic blood pressure (SBP) and pulse before and after RSI. The primary outcomes were change in heart rate and SBP before and after RSI. RESULTS: There were 2,092 patients who met criteria, 85% received etomidate (E), 8% ketamine (K), and 7% propofol (P). Before RSI the ketamine group had a lower SBP (E:135 vs. K:125 vs. P:135 mmHg, p=0.04) but there was no difference in pulse (E:104 vs. K:107 vs. P:105 bpm, p=0.45). After RSI, there were no differences in SBP (E:135 vs. K:130 vs. P:133 mmHg, p=0.34) or pulse (E:106 vs. K:110 vs. P:104 bpm, p=0.08). There was no difference in the average change of SBP (E:0.2 vs. K:5.2 vs. P:-1.8 mmHg, p=0.4) or pulse (E:1.7 vs. K:3.5 bpm vs. P:-0.96, p=0.24) during RSI. CONCLUSION: Contrary to our hypothesis, there was no difference in the hemodynamic effect for etomidate vs. ketamine vs. propofol during RSI in trauma patients. LEVEL OF EVIDENCE: IVStudy TypeRetrospective comparative study without negative criterion.


Posted March 16th 2021

How Should We Sequence the Treatments for Heart Failure and a Reduced Ejection Fraction?: A Redefinition of Evidence-Based Medicine.

Milton Packer M.D.

Milton Packer M.D.

McMurray, J.J.V. and Packer, M. (2021). “How Should We Sequence the Treatments for Heart Failure and a Reduced Ejection Fraction?: A Redefinition of Evidence-Based Medicine.” Circulation 143(9): 875-877.

Full text of this article.

Large-scale trials have demonstrated the efficacy of sacubitril/valsartan, β-blockers, mineralocorticoid receptor antagonists (MRAs), and sodium-glucose cotransporter 2 inhibitors (SGLT2is) as disease-modifying agents that (when combined) represent foundational therapy for heart failure and a reduced ejection fraction (HFrEF). The conventional approach to achieving treatment with all 4 drug classes is to prescribe them in the precise sequence in which they were tested in clinical trials over the past 40 years. Physicians are asked to start with angiotensin-converting enzyme inhibitor or angiotensin receptor blocker, to be followed by a β-blocker, then an MRA, then a neprilysin inhibitor, and, finally, a SGLT2i. Prescribers are advised to titrate the dose of each drug class to the target dose used in large-scale trials before initiating the next recommended drug class. This approach recapitulates the sequence by which these agents were developed for the treatment of heart failure.[No abstract; excerpt from article].


Posted March 16th 2021

Natriuretic Peptide Plasma Concentrations and Risk of Cardiovascular Versus Non-Cardiovascular Events in Heart Failure with Reduced Ejection Fraction Insights from the PARADIGM-HF and ATMOSPHERE Trials.

Milton Packer M.D.

Milton Packer M.D.

Shahzeb Khan, M., Kristensen, S.L., Vaduganathan, M., Kober, L., Abraham, W.T., Desai, A.S., Solomon, S.D., Swedberg, K., Dickstein, K., Zile, M.R., Packer, M., McMurray, J.J. and Butler, J. (2021). “Natriuretic Peptide Plasma Concentrations and Risk of Cardiovascular Versus Non-Cardiovascular Events in Heart Failure with Reduced Ejection Fraction Insights from the PARADIGM-HF and ATMOSPHERE Trials.” Am Heart J Feb 20;S0002-8703(21)00054-5. [Epub ahead of print].

Full text of this article.

BACKGROUND: N-terminal pro-B-type natriuretic peptide (NTproBNP) plasma concentrations are independent prognostic markers in patients with heart failure with reduced ejection fraction (HFrEF). Whether a differential risk association between NTproBNP plasma concentrations and risk of cardiovascular (CV) versus non-CV adverse events exists is not well known. OBJECTIVE: To assess if there is a differential proportional risk of CV versus non CV adverse events by NTproBNP plasma concentrations. METHODS: In this post hoc combined analysis of PARADIGM-HF and ATMOSPHERE trials, proportion of CV versus non-CV mortality and hospitalizations were assessed by NT-proBNP levels (<400, 400-999, 1000-1999, 2000-2999, and >3000 pg/ml) at baseline using Cox regression adjusting for traditional risk factors. RESULTS: 14,737 patients with mean age of 62±8 years (24% history of atrial fibrillation [AF]) were studied. For CV deaths, the event rates per 1000 patient-years steeply increased from 33.8 in the ≤400 pg/ml group to 142.3 in the ≥3000 pg/ml group, while the non-CV death event rates modestly increased from 9.0 to 22.7, respectively. Proportion of non CV deaths decreased across the 5 NT-proBNP groups (21.1%, 18.4%, 17.9%, 17.4%, and 13.7% respectively). Similar trend was observed for non-CV hospitalizations (46.4%, 42.6%, 42.9%, 42.0%, and 36.9% respectively). These results remained similar when stratified according to presence of AF at baseline and prior HF hospitalization within last 12 months. CONCLUSION: The absolute CV event rates per patient years of follow up were greater and had higher stepwise increases than non-CV event rates across a broad range of NT-proBNP plasma concentrations indicating a differential risk of CV events at varying baseline NT-proBNP values. These results have implications for future design of clinical trials.