Deepak Kaji

Deepak Kaji

Auto-generated profile, based on publicly available information

Deepak Kaji

New York, NY, USA

Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
13 February 2019 in PLOS ONE
Verified
Journal Article
Open access
PDF file
DKDeepak A. Kaji
Deepak A. Kaji
JZJohn R. Zech
John R. Zech
JKJun S. Kim
Jun S. Kim
SCSamuel K. Cho
Samuel K. Cho
NDNeha S. Dangayach
Neha S. Dangayach
EOEric K. Oermann
Eric K. Oermann
This study trained long short-term memory (LSTM) recurrent neural networks (RNNs) incorporating an attention mechanism to predict daily sepsis, myocardial infarction (MI), and vancomycin antibiotic administration over two week patient ICU courses in the MIMIC-III dataset. These models achieved next-day predictive AUC of 0.876 for sepsis, 0.823 for MI, and 0.833 for vancomycin administration. Attention maps built from these models highlighted those times when input variables most influenced predictions and could provide a degree of interpretability to clinicians. These models appeared to attend to variables that were proxies for clinician decision-making, demonstrating a challenge of using flexible deep learning approaches trained with EHR data to build clinical decision support. While continued development and refinement is needed, we believe that such models could one day prove useful in reducing information overload for ICU physicians by providing needed clinical decision support for a variety of clinically important tasks.

15 June 2018 in Spine
Verified
Journal Article
Closed access
JKJun S. Kim
Jun S. Kim
RMRobert K. Bs Merrill
Robert K. Bs Merrill
VAVarun Bs Arvind
Varun Bs Arvind
DKDeepak Ba Kaji
Deepak Ba Kaji
SPSara D. Ba Pasik
Sara D. Ba Pasik
SCSamuel K. Cho
Samuel K. Cho
Study Design.: A cross-sectional database study. Objective.: The aim of this study was to train and validate machine learning models to identify risk factors for complications following posterior lumbar spine fusion. Summary of Background Data.: Machine learning models such as artificial neural networks (ANNs) are valuable tools for analyzing and interpreting large and complex datasets. ANNs have yet to be used for risk factor analysis in orthopedic surgery. Methods.: The American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database was queried for patients who underwent posterior lumbar spine fusion. This query returned 22,629 patients, 70% of whom were used to train our models, and 30% were used to evaluate the models. The predictive variables used included sex, age, ethnicity, diabetes, smoking, steroid use, coagulopathy, functional status, American Society for Anesthesiology (ASA) class ≥3, body mass index (BMI), pulmonary comorbidities, and cardiac comorbidities. The models were used to predict cardiac complications, wound complications, venous thromboembolism (VTE), and mortality. Using ASA class as a benchmark for prediction, area under receiver operating curves (AUC) was used to determine the accuracy of our machine learning models. Results.: On the basis of AUC values, ANN and LR both outperformed ASA class for predicting all four types of complications. ANN was the most accurate for predicting cardiac complications, and LR was most accurate for predicting wound complications, VTE, and mortality, though ANN and LR had comparable AUC values for predicting all types of complications. ANN had greater sensitivity than LR for detecting wound complications and mortality. Conclusion.: Machine learning in the form of logistic regression and ANNs were more accurate than benchmark ASA scores for identifying risk factors of developing complications following posterior lumbar spine fusion, suggesting they are potentially great tools for risk factor analysis in spine surgery. Level of Evidence: 3

5 June 2020 in eLife
Verified
Journal Article
Open access
DKDeepak A Kaji
Deepak A Kaji
KHKristen L Howell
Kristen L Howell
ZBZerina Balic
Zerina Balic
DHDirk Hubmacher
Dirk Hubmacher
AHAlice H Huang
Alice H Huang
Tendon injuries are common with poor healing potential. The paucity of therapies for tendon injuries is due to our limited understanding of the cells and molecular pathways that drive tendon regeneration. Using a mouse model of neonatal tendon regeneration, we identified TGFβ signaling as a major molecular pathway that drives neonatal tendon regeneration. Through targeted gene deletion, small molecule inhibition, and lineage tracing, we elucidated TGFβ-dependent and TGFβ-independent mechanisms underlying tendon regeneration. Importantly, functional recovery depended on canonical TGFβ signaling and loss of function is due to impaired tenogenic cell recruitment from both Scleraxis-lineage and non-Scleraxis-lineage sources. We show that TGFβ signaling is directly required in neonatal tenocytes for recruitment and that TGFβ ligand is positively regulated in tendons. Collectively, these results show a functional role for canonical TGFβ signaling in tendon regeneration and offer new insights toward the divergent cellular activities that distinguish regenerative vs fibrotic healing.

1 December 2015 in Nutrition Journal
Verified
Journal Article
Open access
PDF file
RYRabi Yacoub
Rabi Yacoub
DKDeepak Kaji
Deepak Kaji
SPShanti N. Patel
Shanti N. Patel
PSPriya K. Simoes
Priya K. Simoes
DBDeepthi Busayavalasa
Deepthi Busayavalasa
Jaime UribarriJaime Uribarri
Jaime Uribarri
Data from experimental animals suggest that probiotic supplements may retard CKD progression. However, the relationship between probiotic use, frequent yogurt consumption (as a natural probiotic source), and kidney parameters have not been evaluated in humans. We utilized NHANES data, and analyzed the association of probiotic alone (1999–2012) and yogurt/probiotic (2003–2006) use with albuminuria and eGFR after adjustment for demographic and clinical parameters. Frequent yogurt consumption was defined as thrice or more weekly over the year prior to the interview. Frequent yogurt/probiotic consumers had lower adjusted odds of developing combined outcome (albuminuria and/or eGFR < 60 ml/min/1.73 m2) compared to infrequent consumers (OR = 0.76; 95 % CI = 0.61-0.94). When evaluated separately, frequent consumers had lower odds of albuminuria and nonsignificant trend towards decreased odds of low eGFR compared to infrequent consumers. In the probiotic cohort, probiotic consumers were found to have a lower adjusted odds of albuminuria compared to nonusers (OR = 0.59; 95 % CI = 0.37–0.94). Frequent yogurt and/or probiotics use is associated with decreased odds of proteinuric kidney disease. These hypothesis-generating results warrant further translational studies to further delineate the relationship between yogurt/probiotics with kidney dysfunction, as well as microbiome and dysbiosis as potential mediators.

1 June 2018 in Spine
Verified
Journal Article
Closed access
AHAwais K. Ba Hussain
Awais K. Ba Hussain
KVKhushdeep S. Ba Vig
Khushdeep S. Ba Vig
ZCZoe B. Cheung
Zoe B. Cheung
KPKevin Bsc Phan
Kevin Bsc Phan
MLMauricio C. Lima
Mauricio C. Lima
SCSamuel Kang-Wook Cho
Samuel Kang-Wook Cho
Study Design.: A retrospective cohort study from 2011 to 2014 was performed using the American College of Surgeons National Surgical Quality Improvement Program database. Objective.: The purpose of this study was to assess the impact of tumor location in the cervical, thoracic, or lumbosacral spine on 30-day perioperative mortality and morbidity after surgical decompression of metastatic extradural spinal tumors. Summary of Background Data.: Operative treatment of metastatic spinal tumors involves extensive procedures that are associated with significant complication rates and healthcare costs. Past studies have examined various risk factors for poor clinical outcomes after surgical decompression procedures for spinal tumors, but few studies have specifically investigated the impact of tumor location on perioperative mortality and morbidity. Methods.: We identified 2238 patients in the American College of Surgeons National Surgical Quality Improvement Program database who underwent laminectomy for excision of metastatic extradural tumors in the cervical, thoracic, or lumbosacral spine. Baseline patient characteristics were collected from the database. Univariate and multivariate regression analyses were performed to examine the association between spinal tumor location and 30-day perioperative mortality and morbidity. Results.: On univariate analysis, cervical spinal tumors were associated with the highest rate of pulmonary complications. Multivariate regression analysis demonstrated that cervical spinal tumors had the highest odds of multiple perioperative complications. However, thoracic spinal tumors were associated with the highest risk of intra- or postoperative blood transfusion. In contrast, patients with metastatic tumors in the lumbosacral spine had lower odds of perioperative mortality, pulmonary complications, and sepsis. Conclusion.: Tumor location is an independent risk factor for perioperative mortality and morbidity after surgical decompression of metastatic spinal tumors. The addition of tumor location to existing prognostic scoring systems may help to improve their predictive accuracy. Level of Evidence: 3

10 September 2014 in BMC Pulmonary Medicine
Verified
Journal Article
Open access
PDF file
DKDeepak A Kaji
Deepak A Kaji
ABAndrew J Belli
Andrew J Belli
MMMeredith C. McCormack
Meredith C. McCormack
EMElizabeth C Matsui
Elizabeth C Matsui
DWD’ann L Williams
D’ann L Williams
NHNadia Hansel
Nadia Hansel
Indoor particulate matter (PM) has been linked to respiratory symptoms in former smokers with COPD. While subjects with COPD and atopy have also been shown to have more frequent respiratory symptoms, whether they exhibit increased susceptibility to PM as compared to their non-atopic counterparts remains unclear. The aim of this study was to determine whether atopic individuals with COPD have greater susceptibility to PM compared to non-atopic individuals with COPD. Former smokers with moderate to severe COPD were enrolled (n = 77). PM2.5, PM with diameter <2.5 micrometers, was measured in the main living area over three one-week monitoring periods at baseline, 3, and 6 months. Quality of life, respiratory symptoms and medication use were assessed by questionnaires. Serum was analyzed for specific IgE for mouse, cockroach, cat, dog and dust mite allergens. Atopy was established if at least one test was positive. Interaction terms between PM and atopy were tested and generalized estimating equation analysis determined the effect of PM concentrations on health outcomes. Multivariate models were adjusted for age, sex, education, race, season, and baseline lung function and stratified by atopic status. Among atopic individuals, each 10 μg/m3 increase in PM was associated with higher risk of nocturnal symptoms (OR, 1.95; P = 0.02), frequent wheezing (OR, 2.49; P = 0.02), increased rescue medication use (β = 0.14; P = 0.02), dyspnea (β = 0.23; P < 0.001), higher St. George’s Respiratory Quality of Life score (β = 2.55; P = 0.01), and higher breathlessness, cough, and sputum score (BCSS) (β = 0.44; P = 0.01). There was no association between PM and health outcomes among the non-atopic individuals. Interaction terms between PM2.5 and atopy were statistically significant for nocturnal symptoms, frequency of rescue medication use, and BCSS (all P < 0.1). Individuals with COPD and atopy appear to be at higher risk of adverse respiratory health effects of PM exposure compared to non-atopic individuals with COPD.