The Royal South Hants Hospital, known locally as "The RSH", is a community hospital in Southampton. It is managed by NHS Property Services.
Background: Access to vertical and transverse sections of a punch biopsy specimen improves the diagnosis of alopecia. Both two biopsy specimen and single-punch biopsy specimen techniques to visualize both transverse and vertical sections have been described. Their comparative diagnostic certainty is not known. We aimed to assess the diagnostic certainty of a modified HoVert (mHoVert) method, without direct immunofluorescence (DIF), compared to the St John's protocol, a two-biopsy technique with DIF. Methods: Fifty-seven cases of alopecia processed using the St John's protocol and 60 cases of alopecia processed using mHoVert were reviewed. Diagnoses made were rated as certain/probable, possible, or uncertain, depending on the language in the histopathology report. Cases processed by the St John's protocol had final diagnosis and DIF result recorded. Results: In the mHoVert group, significantly more diagnoses were certain/probable (66%, 95% confidence interval [CI]: 57%-75%), compared to 46% (95% CI: 36%-56%) of diagnoses in the St John's protocol group (p = 0.005). DIF result did not affect the final diagnosis in any of the 57 cases reviewed. Conclusions: DIF is not required in the diagnosis of most cases of alopecia. The mHoVert method provides more certain/probable diagnoses than the St John's protocol and can reduce cost and patient morbidity.
Chat Generative Pre-trained Transformer (Chat GPT) is an online language-based platform, designed to answer questions in a human-like way, using deep learning technology.To examine the diagnostic capabilities of Chat GPT using real world, anonymised, medical dermatology cases.Clinical information from 90 consecutive patients referred to a single dermatology emergency clinic between June to December 2022 were examined. Thirty-six patients were included. Anonymised clinical information was transcribed and input into Chat GPT 4.0 followed by the question "what is the most likely diagnosis?". The suggested diagnosis made by Chat GPT was then compared to the diagnosis made in dermatology.After inputting a clinical history and examination by a dermatologist, Chat GPT made a correct primary diagnosis 56% of the time (20/36). Using the clinical history and cutaneous signs from non-specialists, it was able to make a correct diagnosis 38% of the time (14/36). This was similar to the diagnostic rate of non-specialists 36% (13/36), although much lower than dermatologists (83%, 30/36). There was no differential offered by referring sources 27% of the time (10/36), unlike chat GPT which provided a differential diagnosis 100% of the time. Qualitative analysis showed Chat GPT offered responses with caution, often justifying reasoning.This study illustrates that whilst Chat GPT has a diagnostic capability, in its current form it does not significantly improve diagnostic yield in primary or secondary care.
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Dear Editor, We read with interest the article by O'Connor et al. published in a recent issue of Clinical and Experimental Dermatology.1 The authors randomized patients referred by general practitioners (GPs) with suspicious pigmented lesions, to full skin examination (FSE) or targeted lesion examination (TLE) in a 2 : 1 ratio. They reported that just 7.1% (n = 37) of patients had incidental lesions detected by FSE, including one incidental skin cancer, which was a superficial basal cell carcinoma (BCC). We report our findings from a study of a skin cancer screening clinic. We undertook a service evaluation (SE) on a high-volume skin cancer screening clinic, which offers FSEs to all patients. The study was granted ethics approval by the Faculty of Medicine Ethics Committee of the University of Southampton. Patients were referred by their GP on a 2-week-wait pathway for suspected skin cancer. We found that 73.2% (n = 120) of patients who had an FSE or partial skin examination had one or more lesions detected incidentally. The diagnoses of all referred index skin cancers (n = 60) and incidental skin cancers (n = 27) are summarized in Table 1. The most common skin cancer diagnosis was BCC, with 40 index lesions plus an additional 24 detected incidentally. In our study, 13 malignant melanomas (MMs) were diagnosed overall, of which 9 were index lesions. Two of the index MMs were melanoma in situ (MIS), and the mean ± SD Breslow depth of the seven invasive index MMs was 1.94 ± 1.85 mm. Four additional incidental MMs were identified via FSE (Table 2), of which one was an MIS; the Breslow depth of the three invasive incidental MMs was 1.03 ± 0.50 mm. All incidental MMs were located in a site distant from the referred index lesion, highlighting that these lesions would not have been detected without FSE. Intriguingly, however, all would have been detected by a partial examination of the head, neck and trunk. We found that 30% of invasive MMs were detected incidentally and surmise that had these not been detected at FSE, they would most likely have presented at a later stage with a thicker Breslow depth and worse prognosis requiring more complex and costly investigation and treatment. We agree that the COVID-19 pandemic significantly increased pressure on dermatology resources. Unfortunately, patients' access to dermatology practitioners was reduced and consequently, there was a substantial decrease in the number of skin cancers diagnosed and treated.2 We would argue that it is our duty to help our inundated primary care colleagues by pre-emptive screening partial or FSEs to prevent late disease presentation, which is more costly to treat and is associated with poorer patient outcomes. Our results may differ from those of O'Connor et al. due to different populations and healthcare systems, but they seem curiously markedly at odds with those, and are more in line with the results published by both Aldridge et al.,3 who noted 30% incident melanomas, and Jiang et al.,4 who found markedly lower Breslow depth in incidentally detected MMs. Based on these published findings3, 4 and our SE, our adage will continue follow the adage of ‘Seek and ye shall find’ until we can be convinced that this is not in the best interests of patients.
