Ginikanwa Onyekaba

Bio: Ginikanwa Onyekaba is an academic researcher from University of Pennsylvania. The author has contributed to research in topics: Medicine & Central centrifugal cicatricial alopecia. The author has an hindex of 2, co-authored 8 publications receiving 48 citations.

Characterization of the inflammatory features of central centrifugal cicatricial alopecia.

Abstract: Central centrifugal cicatricial alopecia (CCCA) is a scarring alopecia that primarily affects women of African descent. Although histopathological features of CCCA have been described, the pathophysiology of this disease remains unclear. To better understand the components of CCCA pathophysiology, we evaluated the composition of the inflammatory infiltrate, the distribution of Langerhans cells (LCs), and the relationship between fibrosis and perifollicular vessel distribution. Our data indicate that CCCA is associated with a CD4-predominant T-cell infiltrate with increased LCs extending into the lower hair follicle. Fibroplasia associated with follicular scarring displaces blood vessels away from the outer root sheath epithelium. These data indicate that CCCA is an inflammatory scarring alopecia with unique pathophysiologic features that differentiate it from other lymphocytic scarring processes.

Does Dedicated Research Time During Residency Predict Resident Academic Productivity and a Future Academic Career?

Abstract: Objective Dedicated research time is a component of certain plastic surgery programs, and yet, there is limited research examining its impact on academic productivity and career outcomes. This study aimed to assess the effect of dedicated research time on the academic productivity of residents and the likelihood of pursuing an academic career. Methods We conducted a cross-sectional study that examined bibliometric indices of integrated plastic surgery residency graduates from 2010 to 2020. Academic productivity was determined by the number of peer-reviewed publications and h-index 1 year after residency graduation. Results were analyzed using descriptive statistics, χ2 test, t test, and logistic regression. Results Data on plastic surgery residency graduates were analyzed (N = 490 from 46 programs). The mean numbers of publications and h-index per research track graduate were 26.1 and 8.23, respectively. The mean numbers of publications and h-index per nonresearch track graduate were 15.9 and 5.97, respectively. After controlling for the University of Alabama research ranking through multilinear regression analysis, we found that pursuing dedicated research time was an independent predictor of increased h-index and publication output, although it did not predict an increased likelihood of pursuing an academic career. Conclusions Participating in dedicated research during residency increases academic productivity, irrespective of the residency program’s research rank. Given this finding, offering research years can help support the mission of fostering academic opportunities within plastic surgery.

Comorbid anxiety and depression among black women with central centrifugal cicatricial alopecia: A retrospective study.

Abstract: Dear Editor, The published work suggests a higher risk for anxiety and depression among patients with alopecia areata (AA), but less is known about central centrifugal cicatricial alopecia (CCCA), a primary scarring alopecia. CCCA occurs predominantly in black women, with a prevalence of 2.7–5.7%. In a previous study, CCCA was strongly associated with low quality of life. However, the psychiatric comorbidity burden among patients with CCCA has not been explored. We aimed to assess the risk of comorbid depression and anxiety among black female patients with CCCA compared with patients with AA and psoriasis. In this cross-sectional study, the medical records of black female patients aged 18 years and older with a diagnosis of CCCA, psoriasis or alopecia areata seen in the Dermatology Department of Perelman School of Medicine between July 2017 and July 2019 were reviewed. This study received an exemption from the institutional review board of the University of Pennsylvania. International Classification of Diseases, 10th Revision codes included L40.0 (psoriasis vulgaris), L66.9 (CCCA), L63.9 (AA), F32.1 (major depressive disorder, single episode, moderate), F32.2 (major depressive disorder, single episode), F33.0 (major depressive disorder, recurrent, mild), F33.1 (major depressive disorder, recurrent, moderate), F32.2 (major depressive disorder, recurrent, severe) and F41.1 (generalized anxiety disorder). Data was collected on patient demographics, and patients with a concomitant diagnosis of anxiety or depression were identified. Descriptive statistics were calculated, P-values were calculated using a two-tailed v-test with Yates’ correction, and odds ratios were calculated with 95% confidence intervals by Stata version 16 software (StataCorp, College Station, TX, USA). There were a total of 270 black women with a diagnosis of CCCA. Of these, 27 (10.0%) had a diagnosis of anxiety or depression in their medical records. In the same period, 84 black women with psoriasis and 69 black women with AA were identified. In these groups, anxiety or depression was accounted for in 10 (11.9%) patients with psoriasis and seven (10.1%) with AA. No significant difference was found in the risk for anxiety or depression in CCCA compared with psoriasis (P = 0.68) and AA (P = 0.84) (Table 1). In this study, the risk of psychiatric comorbidity for patients with CCCA was not significantly different than that of psoriasis and AA. This may represent a clinically significant observation. Presently, treatment options for CCCA rarely achieve hair regrowth, and halting the progression of disease is the main goal. Additionally, there is a disparity in access to mental health care among black patients, making the CCCA population particularly vulnerable. Patients with CCCA may greatly benefit from psychological screening and interventions. Limitations of this study include its retrospective nature, small cohort size and single academic center setting with a special interest in scarring alopecia, explaining the high number of patients with scarring alopecia seen in this study. Further investigation is required to better characterize the psychiatric comorbidity burden of CCCA. This vital information could better address the needs of this patient population.

Underspecification Presents Challenges for Credibility in Modern Machine Learning

Abstract: ML models often exhibit unexpectedly poor behavior when they are deployed in real-world domains. We identify underspecification as a key reason for these failures. An ML pipeline is underspecified when it can return many predictors with equivalently strong held-out performance in the training domain. Underspecification is common in modern ML pipelines, such as those based on deep learning. Predictors returned by underspecified pipelines are often treated as equivalent based on their training domain performance, but we show here that such predictors can behave very differently in deployment domains. This ambiguity can lead to instability and poor model behavior in practice, and is a distinct failure mode from previously identified issues arising from structural mismatch between training and deployment domains. We show that this problem appears in a wide variety of practical ML pipelines, using examples from computer vision, medical imaging, natural language processing, clinical risk prediction based on electronic health records, and medical genomics. Our results show the need to explicitly account for underspecification in modeling pipelines that are intended for real-world deployment in any domain.

Evaluating Deep Neural Networks Trained on Clinical Images in Dermatology with the Fitzpatrick 17k Dataset

Abstract: How does the accuracy of deep neural network models trained to classify clinical images of skin conditions vary across skin color? While recent studies demonstrate computer vision models can serve as a useful decision support tool in healthcare and provide dermatologist-level classification on a number of specific tasks, darker skin is under-represented in the data. Most publicly available data sets do not include Fitzpatrick skin type labels. We annotate 16,577 clinical images sourced from two dermatology atlases with Fitzpatrick skin type labels and open-source these annotations. Based on these labels, we find that there are significantly more images of light skin types than dark skin types in this dataset. We train a deep neural network model to classify 114 skin conditions and find that the model is most accurate on skin types similar to those it was trained on. In addition, we evaluate how an algorithmic approach to identifying skin tones, individual typology angle, compares with Fitzpatrick skin type labels annotated by a team of human labelers.

Common causes of hair loss ‐ clinical manifestations, trichoscopy and therapy

Abstract: The field of hair disorders is constantly growing. The most important hair diseases are divided in non- cicatricial and cicatricial ones. Non-cicatricial alopecia are more frequent than cicatricial alopecia. The first step is to obtain a good history and physical examination. Laboratory testing is often unnecessary, while trichoscopy is fundamental for all hair diseases. Scalp biopsy is strongly suggested in cicatricial alopecia and in doubtful cases. Androgenetic alopecia, alopecia areata, telogen effluvium, trichotillomania are common causes of non- cicatricial alopecia. Frontal fibrosing alopecia, discoid lupus erythematosus, lichen planopilaris, follicullitis decalvans are some of the most common forms of cicatricial hair loss. Many treatments are available, and a prompt diagnosis is very important for the prognosis.

Disparities in dermatology AI performance on a diverse, curated clinical image set

Abstract: An estimated 3 billion people lack access to dermatological care globally. Artificial intelligence (AI) may aid in triaging skin diseases and identifying malignancies. However, most AI models have not been assessed on images of diverse skin tones or uncommon diseases. Thus, we created the Diverse Dermatology Images (DDI) dataset—the first publicly available, expertly curated, and pathologically confirmed image dataset with diverse skin tones. We show that state-of-the-art dermatology AI models exhibit substantial limitations on the DDI dataset, particularly on dark skin tones and uncommon diseases. We find that dermatologists, who often label AI datasets, also perform worse on images of dark skin tones and uncommon diseases. Fine-tuning AI models on the DDI images closes the performance gap between light and dark skin tones. These findings identify important weaknesses and biases in dermatology AI that should be addressed for reliable application to diverse patients and diseases.