Kristen W. Yeom

Kristen W. Yeom

Auto-generated profile, based on publicly available information

Kristen W. Yeom

Department of Neurosurgery, Stanford University School of Medicine, Stanford, USA

Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

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

Representative Publications

View All Publications
20 November 2018 in PLOS Medicine
Verified
Journal Article
Open access
PDF file
PRPranav Rajpurkar
Pranav Rajpurkar
JIJeremy Irvin
Jeremy Irvin
RBRobyn L Ball
Robyn L Ball
KZKaylie Zhu
Kaylie Zhu
BYBrandon Yang
Brandon Yang
MLMatthew P Lungren
Matthew P Lungren
Chest radiograph interpretation is critical for the detection of thoracic diseases, including tuberculosis and lung cancer, which affect millions of people worldwide each year. This time-consuming task typically requires expert radiologists to read the images, leading to fatigue-based diagnostic error and lack of diagnostic expertise in areas of the world where radiologists are not available. Recently, deep learning approaches have been able to achieve expert-level performance in medical image interpretation tasks, powered by large network architectures and fueled by the emergence of large labeled datasets. The purpose of this study is to investigate the performance of a deep learning algorithm on the detection of pathologies in chest radiographs compared with practicing radiologists. We developed CheXNeXt, a convolutional neural network to concurrently detect the presence of 14 different pathologies, including pneumonia, pleural effusion, pulmonary masses, and nodules in frontal-view chest radiographs. CheXNeXt was trained and internally validated on the ChestX-ray8 dataset, with a held-out validation set consisting of 420 images, sampled to contain at least 50 cases of each of the original pathology labels. On this validation set, the majority vote of a panel of 3 board-certified cardiothoracic specialist radiologists served as reference standard. We compared CheXNeXt’s discriminative performance on the validation set to the performance of 9 radiologists using the area under the receiver operating characteristic curve (AUC). The radiologists included 6 board-certified radiologists (average experience 12 years, range 4–28 years) and 3 senior radiology residents, from 3 academic institutions. We found that CheXNeXt achieved radiologist-level performance on 11 pathologies and did not achieve radiologist-level performance on 3 pathologies. The radiologists achieved statistically significantly higher AUC performance on cardiomegaly, emphysema, and hiatal hernia, with AUCs of 0.888 (95% confidence interval [CI] 0.863–0.910), 0.911 (95% CI 0.866–0.947), and 0.985 (95% CI 0.974–0.991), respectively, whereas CheXNeXt’s AUCs were 0.831 (95% CI 0.790–0.870), 0.704 (95% CI 0.567–0.833), and 0.851 (95% CI 0.785–0.909), respectively. CheXNeXt performed better than radiologists in detecting atelectasis, with an AUC of 0.862 (95% CI 0.825–0.895), statistically significantly higher than radiologists' AUC of 0.808 (95% CI 0.777–0.838); there were no statistically significant differences in AUCs for the other 10 pathologies. The average time to interpret the 420 images in the validation set was substantially longer for the radiologists (240 minutes) than for CheXNeXt (1.5 minutes). The main limitations of our study are that neither CheXNeXt nor the radiologists were permitted to use patient history or review prior examinations and that evaluation was limited to a dataset from a single institution. In this study, we developed and validated a deep learning algorithm that classified clinically important abnormalities in chest radiographs at a performance level comparable to practicing radiologists. Once tested prospectively in clinical settings, the algorithm could have the potential to expand patient access to chest radiograph diagnostics.

7 February 2022 in Nature
Verified
Journal Article
Open access
PDF file
RMRobbie G. Majzner
Robbie G. Majzner
SRSneha Ramakrishna
Sneha Ramakrishna
KYKristen W. Yeom
Kristen W. Yeom
SPShabnum Patel
Shabnum Patel
HCHarshini Chinnasamy
Harshini Chinnasamy
MMMichelle Monje
Michelle Monje
Diffuse intrinsic pontine glioma (DIPG) and other H3K27M-mutated diffuse midline gliomas (DMGs) are universally lethal paediatric tumours of the central nervous system1. We have previously shown that the disialoganglioside GD2 is highly expressed on H3K27M-mutated glioma cells and have demonstrated promising preclinical efficacy of GD2-directed chimeric antigen receptor (CAR) T cells2, providing the rationale for a first-in-human phase I clinical trial (NCT04196413). Because CAR T cell-induced brainstem inflammation can result in obstructive hydrocephalus, increased intracranial pressure and dangerous tissue shifts, neurocritical care precautions were incorporated. Here we present the clinical experience from the first four patients with H3K27M-mutated DIPG or spinal cord DMG treated with GD2-CAR T cells at dose level 1 (1 × 106 GD2-CAR T cells per kg administered intravenously). Patients who exhibited clinical benefit were eligible for subsequent GD2-CAR T cell infusions administered intracerebroventricularly3. Toxicity was largely related to the location of the tumour and was reversible with intensive supportive care. On-target, off-tumour toxicity was not observed. Three of four patients exhibited clinical and radiographic improvement. Pro-inflammatory cytokine levels were increased in the plasma and cerebrospinal fluid. Transcriptomic analyses of 65,598 single cells from CAR T cell products and cerebrospinal fluid elucidate heterogeneity in response between participants and administration routes. These early results underscore the promise of this therapeutic approach for patients with H3K27M-mutated DIPG or spinal cord DMG.

27 November 2018 in PLOS Medicine
Verified
Journal Article
Open access
PDF file
NBNicholas Bien
Nicholas Bien
PRPranav Rajpurkar
Pranav Rajpurkar
RBRobyn L Ball
Robyn L Ball
JIJeremy Irvin
Jeremy Irvin
APAllison Park
Allison Park
MLMatthew P Lungren
Matthew P Lungren
Magnetic resonance imaging (MRI) of the knee is the preferred method for diagnosing knee injuries. However, interpretation of knee MRI is time-intensive and subject to diagnostic error and variability. An automated system for interpreting knee MRI could prioritize high-risk patients and assist clinicians in making diagnoses. Deep learning methods, in being able to automatically learn layers of features, are well suited for modeling the complex relationships between medical images and their interpretations. In this study we developed a deep learning model for detecting general abnormalities and specific diagnoses (anterior cruciate ligament [ACL] tears and meniscal tears) on knee MRI exams. We then measured the effect of providing the model’s predictions to clinical experts during interpretation. Our dataset consisted of 1,370 knee MRI exams performed at Stanford University Medical Center between January 1, 2001, and December 31, 2012 (mean age 38.0 years; 569 [41.5%] female patients). The majority vote of 3 musculoskeletal radiologists established reference standard labels on an internal validation set of 120 exams. We developed MRNet, a convolutional neural network for classifying MRI series and combined predictions from 3 series per exam using logistic regression. In detecting abnormalities, ACL tears, and meniscal tears, this model achieved area under the receiver operating characteristic curve (AUC) values of 0.937 (95% CI 0.895, 0.980), 0.965 (95% CI 0.938, 0.993), and 0.847 (95% CI 0.780, 0.914), respectively, on the internal validation set. We also obtained a public dataset of 917 exams with sagittal T1-weighted series and labels for ACL injury from Clinical Hospital Centre Rijeka, Croatia. On the external validation set of 183 exams, the MRNet trained on Stanford sagittal T2-weighted series achieved an AUC of 0.824 (95% CI 0.757, 0.892) in the detection of ACL injuries with no additional training, while an MRNet trained on the rest of the external data achieved an AUC of 0.911 (95% CI 0.864, 0.958). We additionally measured the specificity, sensitivity, and accuracy of 9 clinical experts (7 board-certified general radiologists and 2 orthopedic surgeons) on the internal validation set both with and without model assistance. Using a 2-sided Pearson’s chi-squared test with adjustment for multiple comparisons, we found no significant differences between the performance of the model and that of unassisted general radiologists in detecting abnormalities. General radiologists achieved significantly higher sensitivity in detecting ACL tears (p-value = 0.002; q-value = 0.019) and significantly higher specificity in detecting meniscal tears (p-value = 0.003; q-value = 0.019). Using a 1-tailed t test on the change in performance metrics, we found that providing model predictions significantly increased clinical experts’ specificity in identifying ACL tears (p-value < 0.001; q-value = 0.006). The primary limitations of our study include lack of surgical ground truth and the small size of the panel of clinical experts. Our deep learning model can rapidly generate accurate clinical pathology classifications of knee MRI exams from both internal and external datasets. Moreover, our results support the assertion that deep learning models can improve the performance of clinical experts during medical imaging interpretation. Further research is needed to validate the model prospectively and to determine its utility in the clinical setting.

5 October 2017 in American Journal of Neuroradiology
Verified
Journal Article
Open access
PDF file
MZM. Zhou
M. Zhou
JSJacob G Scott
Jacob G Scott
BCB. Chaudhury
B. Chaudhury
LHLawrence Hall
Lawrence Hall
Dmitry GoldgofDmitry Goldgof
Dmitry Goldgof
RGRobert Gatenby
Robert Gatenby
SUMMARY: Radiomics describes a broad set of computational methods that extract quantitative features from radiographic images. The resulting features can be used to inform imaging diagnosis, prognosis, and therapy response in oncology. However, major challenges remain for methodologic developments to optimize feature extraction and provide rapid information flow in clinical settings. Equally important, to be clinically useful, predictive radiomic properties must be clearly linked to meaningful biologic characteristics and qualitative imaging properties familiar to radiologists. Here we use a cross-disciplinary approach to highlight studies in radiomics. We review brain tumor radiologic studies (eg, imaging interpretation) through computational models (eg, computer vision and machine learning) that provide novel clinical insights. We outline current quantitative image feature extraction and prediction strategies with different levels of available clinical classes for supporting clinical decision-making. We further discuss machine-learning challenges and data opportunities to advance radiomic studies.

15 May 2014 in American Journal of Neuroradiology
Verified
Journal Article
Open access
PDF file
SPS. Perreault
S. Perreault
Vijay RamaswamyVijay Ramaswamy
Vijay Ramaswamy
AAA.S. Achrol
A.S. Achrol
KCK. Chao
K. Chao
TLT.T. Liu
T.T. Liu
KYK.W. Yeom
K.W. Yeom
BACKGROUND AND PURPOSE: Recently identified molecular subgroups of medulloblastoma have shown potential for improved risk stratification. We hypothesized that distinct MR imaging features can predict these subgroups. MATERIALS AND METHODS: All patients with a diagnosis of medulloblastoma at one institution, with both pretherapy MR imaging and surgical tissue, served as the discovery cohort (n = 47). MR imaging features were assessed by 3 blinded neuroradiologists. NanoString-based assay of tumor tissues was conducted to classify the tumors into the 4 established molecular subgroups (wingless, sonic hedgehog, group 3, and group 4). A second pediatric medulloblastoma cohort (n = 52) from an independent institution was used for validation of the MR imaging features predictive of the molecular subtypes. RESULTS: Logistic regression analysis within the discovery cohort revealed tumor location (P < .001) and enhancement pattern (P = .001) to be significant predictors of medulloblastoma subgroups. Stereospecific computational analyses confirmed that group 3 and 4 tumors predominated within the midline fourth ventricle (100%, P = .007), wingless tumors were localized to the cerebellar peduncle/cerebellopontine angle cistern with a positive predictive value of 100% (95% CI, 30%–100%), and sonic hedgehog tumors arose in the cerebellar hemispheres with a positive predictive value of 100% (95% CI, 59%–100%). Midline group 4 tumors presented with minimal/no enhancement with a positive predictive value of 91% (95% CI, 59%–98%). When we used the MR imaging feature–based regression model, 66% of medulloblastomas were correctly predicted in the discovery cohort, and 65%, in the validation cohort. CONCLUSIONS: Tumor location and enhancement pattern were predictive of molecular subgroups of pediatric medulloblastoma and may potentially serve as a surrogate for genomic testing.

5 June 2019 in JAMA Network Open
Verified
Journal Article
Open access
PDF file
APAllison Park
Allison Park
CCChris Chute
Chris Chute
PRPranav Rajpurkar
Pranav Rajpurkar
JLJoe Lou
Joe Lou
RBRobyn L. Ball
Robyn L. Ball
KYKristen W. Yeom
Kristen W. Yeom
Diagnosis of unruptured aneurysms is a critically important clinical task: intracranial aneurysms occur in 1% to 3% of the population and account for more than 80% of nontraumatic life-threatening subarachnoid hemorrhages.1 Computed tomographic angiography (CTA) is the primary, minimally invasive imaging modality currently used for diagnosis, surveillance, and presurgical planning of intracranial aneurysms,2,3 but interpretation is time consuming even for subspecialty-trained neuroradiologists. Low interrater agreement poses an additional challenge for reliable diagnosis.4-7