Evaluating Artificial Intelligence's Role in Teaching the Reporting and Interpretation of Computed Tomographic Angiography for Preoperative Planning of the Deep Inferior Epigastric Artery Perforator Flap.

Background: Artificial intelligence (AI) has the potential to transform preoperative planning for breast reconstruction by enhancing the efficiency, accuracy, and reliability of radiology reporting through automatic interpretation and perforator identification. Large language models (LLMs) have recently advanced significantly in medicine. This study aimed to evaluate the proficiency of contemporary LLMs in interpreting computed tomography angiography (CTA) scans for deep inferior epigastric perforator (DIEP) flap preoperative planning. Methods: Four prominent LLMs, ChatGPT-4, BARD, Perplexity, and BingAI, answered six questions on CTA scan reporting. A panel of expert plastic surgeons with extensive experience in breast reconstruction assessed the responses using a Likert scale. In contrast, the responses' readability was evaluated using the Flesch Reading Ease score, the Flesch-Kincaid Grade level, and the Coleman-Liau Index. The DISCERN score was utilized to determine the responses' suitability. Statistical significance was identified through a t-test, and P-values < 0.05 were considered significant. Results: BingAI provided the most accurate and useful responses to prompts, followed by Perplexity, ChatGPT, and then BARD. BingAI had the greatest Flesh Reading Ease (34.7±5.5) and DISCERN (60.5±3.9) scores. Perplexity had higher Flesch-Kincaid Grade level (20.5±2.7) and Coleman-Liau Index (17.8±1.6) scores than other LLMs. Conclusion: LLMs exhibit limitations in their capabilities of reporting CTA for preoperative planning of breast reconstruction, yet the rapid advancements in technology hint at a promising future. AI stands poised to enhance the education of CTA reporting and aid preoperative planning. In the future, AI technology could provide automatic CTA interpretation, enhancing the efficiency, accuracy, and reliability of CTA reports.

Use of artificial intelligence in breast surgery: a narrative review.

Background and Objective: We have witnessed tremendous advances in artificial intelligence (AI) technologies. Breast surgery, a subspecialty of general surgery, has notably benefited from AI technologies. This review aims to evaluate how AI has been integrated into breast surgery practices, to assess its effectiveness in improving surgical outcomes and operational efficiency, and to identify potential areas for future research and application. Methods: Two authors independently conducted a comprehensive search of PubMed, Google Scholar, EMBASE, and Cochrane CENTRAL databases from January 1, 1950, to September 4, 2023, employing keywords pertinent to AI in conjunction with breast surgery or cancer. The search focused on English language publications, where relevance was determined through meticulous screening of titles, abstracts, and full-texts, followed by an additional review of references within these articles. The review covered a range of studies illustrating the applications of AI in breast surgery encompassing lesion diagnosis to postoperative follow-up. Publications focusing specifically on breast reconstruction were excluded. Key Content and Findings: AI models have preoperative, intraoperative, and postoperative applications in the field of breast surgery. Using breast imaging scans and patient data, AI models have been designed to predict the risk of breast cancer and determine the need for breast cancer surgery. In addition, using breast imaging scans and histopathological slides, models were used for detecting, classifying, segmenting, grading, and staging breast tumors. Preoperative applications included patient education and the display of expected aesthetic outcomes. Models were also designed to provide intraoperative assistance for precise tumor resection and margin status assessment. As well, AI was used to predict postoperative complications, survival, and cancer recurrence. Conclusions: Extra research is required to move AI models from the experimental stage to actual implementation in healthcare. With the rapid evolution of AI, further applications are expected in the coming years including direct performance of breast surgery. Breast surgeons should be updated with the advances in AI applications in breast surgery to provide the best care for their patients.

IL7 in combination with radiotherapy stimulates a memory T-cell response to improve outcomes in HNSCC models.

Clinically approved head and neck squamous cell carcinoma (HNSCC) immunotherapies manipulate the immune checkpoint blockade (ICB) axis but have had limited success outside of recurrent/metastatic disease. Interleukin-7 (IL7) has been shown to be essential for effector T-cell survival, activation, and proliferation. Here, we show that IL7 in combination with radiotherapy (RT) is effective in activating CD8 + T-cells for reducing tumor growth. Our studies were conducted using both human papillomavirus related and unrelated orthotopic HNSCC murine models. Immune populations from the tumor, draining lymph nodes, and blood were compared between treatment groups and controls using flow cytometry, proteomics, immunofluorescence staining, and RNA sequencing. Treatment with RT and IL7 (RT + IL7) resulted in significant tumor growth reduction, high CD8 T-cell tumor infiltration, and increased proliferation of T-cell progenitors in the bone marrow. IL7 also expanded a memory-like subpopulation of CD8 T-cells. These results indicate that IL7 in combination with RT can serve as an effective immunotherapy strategy outside of the conventional ICB axis to drive the antitumor activity of CD8 T-cells.

Artificial Intelligence Language Model Performance for Rapid Intraoperative Queries in Plastic Surgery: ChatGPT and the Deep Inferior Epigastric Perforator Flap.

Background: The integration of artificial intelligence in healthcare has led to the development of large language models that can address various medical queries, including intraoperatively. This study investigates the potential of ChatGPT in addressing intraoperative questions during the deep inferior epigastric perforator flap procedure. Methods: A series of six intraoperative questions specific to the DIEP flap procedure, derived from real-world clinical scenarios, were proposed to ChatGPT. A panel of four experienced board-certified plastic surgeons evaluated ChatGPT's performance in providing accurate, relevant, and comprehensible responses. Results: The Likert scale demonstrated to be medically accurate, systematic in presentation, and logical when providing alternative solutions. The mean readability score of the Flesch Reading Ease Score was 28.7 (±0.8), the Flesch-Kincaid Grade Level was 12.4 (±0.5), and the Coleman-Liau Index was 14.5 (±0.5). Suitability-wise, the DISCERN score of ChatGPT was 48 (±2.5) indicating suitable and comprehensible language for experts. Conclusions: Generative AI tools such as ChatGPT can serve as a supplementary tool for surgeons to offer valuable insights and foster intraoperative problem-solving abilities. However, it lacks consideration of individual patient factors and surgical nuances. Nevertheless, further refinement of its training data and rigorous scrutiny under experts to ensure the accuracy and up-to-date nature of the information holds the potential for it to be utilized in the surgical field.

Sensory nerve release of CGRP increases tumor growth in HNSCC by suppressing TILs.

BACKGROUND: Perineural invasion (PNI) and nerve density within the tumor microenvironment (TME) have long been associated with worse outcomes in head and neck squamous cell carcinoma (HNSCC). This prompted an investigation into how nerves within the tumor microenvironment affect the adaptive immune system and tumor growth. METHODS: We used RNA sequencing analysis of human tumor tissue from a recent HNSCC clinical trial, proteomics of human nerves from HNSCC patients, and syngeneic orthotopic murine models of HPV-unrelated HNSCC to investigate how sensory nerves modulate the adaptive immune system. FINDINGS: Calcitonin gene-related peptide (CGRP) directly inhibited CD8 T cell activity in vitro, and blocking sensory nerve function surgically, pharmacologically, or genetically increased CD8 and CD4 T cell activity in vivo. CONCLUSIONS: Our data support sensory nerves playing a role in accelerating tumor growth by directly acting on the adaptive immune system to decrease Th1 CD4 T cells and activated CD8 T cells in the TME. These data support further investigation into the role of sensory nerves in the TME of HNSCC and points toward the possible treatment efficacy of blocking sensory nerve function or specifically inhibiting CGRP release or activity within the TME to improve outcomes. FUNDING: 1R01DE028282-01, 1R01DE028529-01, 1P50CA261605-01 (to S.D.K.), 1R01CA284651-01 (to S.D.K.), and F31 DE029997 (to L.B.D.).

A retrospective study on weaning glucocorticoids and recovery of the hypothalamic-pituitary-adrenal axis.

CONTEXT: Glucocorticoids suppress the hypothalamic-pituitary-adrenal (HPA) axis resulting in tertiary adrenal insufficiency (AI). When weaning patients off glucocorticoids there is no consensus on whether to maintain patients on prednisolone or convert to hydrocortisone. OBJECTIVE: Investigate HPA axis recovery in patients on long-term prednisolone and assess outcome after hydrocortisone conversion. DESIGN: Retrospective cohort study. SETTING: Outpatient endocrine steroid clinic. PATIENTS: Patients on long-term prednisolone referred for HPA axis testing between 2015-2022. MAIN OUTCOMES MEASURED: 1) HPA axis recovery rate in patients on prednisolone demonstrated by normal ACTH stimulation test (AST).2) HPA axis recovery rate sub-analysis of dose-matched patients with confirmed tertiary AI on prednisolone or hydrocortisone. RESULTS: 206 patients on prednisolone were tested for tertiary AI. Of these 176 remained on prednisolone while 30 were converted to hydrocortisone. The overall HPA axis recovery rate for patients on prednisolone after interval testing was 137/206 (66.5%). HPA axis recovery rate in dose-matched prednisolone and hydrocortisone conversion groups was 7/10 (70%) and 2/13 (15%) (p=0.008), respectively. There was no difference in mean (SD) age (67.1(12.2) v 63.4(11.1) years; p=0.464) and baseline cortisol (5.3(4.2) v 4.6(3.1)µg/dL; p=0.648) and median [IQR] glucocorticoids duration (1213[1114] v 2316[4808] days; p=0.693) and baseline ACTH (20.5[29.0] v 16.3[14.8]ng/L; p=0.905) between dose-matched prednisolone and hydrocortisone groups. Follow-up duration in prednisolone group was significantly lower (median [IQR] 348[975] v 667[884] days; p=0.012). CONCLUSIONS: Patients with glucocorticoid induced AI maintained on once-daily prednisolone can recover HPA axis function when weaning. There is no apparent advantage to recover HPA axis function in converting to multiple dosing hydrocortisone.

PPARα Agonism Enhances Immune Response to Radiotherapy While Dietary Oleic Acid Results in Counteraction.

PURPOSE: Head and neck cancer (HNC) improvements are stagnant, even with advances in immunotherapy. Our previous clinical trial data show that altered fatty acid (FA) metabolism correlates with outcome. We hypothesized that pharmacologic and dietary modulation of FA catabolism will affect therapeutic efficacy. EXPERIMENTAL DESIGN: We performed in vivo and in vitro experiments using PPARα agonism with fenofibrate (FF) or high oleic acid diets (OAD) with radiotherapy, generating metabolomic, proteomic, stable isotope tracing, extracellular flux analysis, and flow-cytometric data to investigate these alterations. RESULTS: FF improved antitumor efficacy of high dose per fraction radiotherapy in HNC murine models, whereas the OAD reversed this effect. FF-treated mice on the control diet had evidence of increased FA catabolism. Stable isotope tracing showed less glycolytic utilization by ex vivo CD8+ T cells. Improved efficacy correlated with intratumoral alterations in eicosanoid metabolism and downregulated mTOR and CD36. CONCLUSIONS: Metabolic intervention with increased FA catabolism improves the efficacy of HNC therapy and enhances antitumoral immune response.

Differential recruitment of brain circuits during fear extinction in non-stressed compared to stress resilient animals.

Dysfunctional fear responses in post-traumatic stress disorder (PTSD) may be partly explained by an inability to effectively extinguish fear responses elicited by trauma-related cues. However, only a subset of individuals exposed to traumatic stress develop PTSD. Therefore, studying fear extinction deficits in animal models of individual differences could help identify neural substrates underlying vulnerability or resilience to the effects of stress. We used a rat model of social defeat in which rats segregate into passively and actively coping rats. In previous work, we showed that passively coping rats exhibit disruptions in social interaction whereas actively coping rats do not display behaviors differently from controls, indicating their resilience. Here, adult male rats exposed to 7 days of social defeat were tested for fear extinction, retention of extinction, and persistence of retention using contextual fear and ethologically-relevant fear tests. Passively coping rats exhibited elevated freezing in response to the previously extinguished context. Analyses of cFos expressing cells across select brain regions showed high correlations within dorsal hippocampal subregions, while passively coping rats had high correlations between the dorsal hippocampus CA1 and the central and basolateral subregions of the amygdala. Importantly, although control and actively coping rats showed similar levels of behavioral extinction, there was little similarity between activated structures, suggesting stress resilience in response to chronic social defeat involves an adaptive differential recruitment of brain circuits to successfully extinguish fear memories.

The importance of REM sleep fragmentation in the effects of stress on sleep: Perspectives from preclinical studies.

Psychological stress poses a risk for sleep disturbances. Importantly, trauma-exposed individuals who develop posttraumatic stress disorder (PTSD) frequently report insomnia and recurrent nightmares. Clinical studies have provided insight into the mechanisms of these sleep disturbances. We review polysomnographic findings in PTSD and identify analogous measures that have been made in animal models of PTSD. There is a rich empirical and theoretical literature on rapid eye movement sleep (REMS) substrates of insomnia and nightmares, with an emphasis on REMS fragmentation. For future investigations of stress-induced sleep changes, we recommend a focus on tonic, phasic and other microarchitectural REMS measures. Power spectral density analysis of the sleep EEG should also be utilized. Animal models with high construct validity can provide insight into gender and time following stressor exposure as moderating variables. Ultimately, preclinical studies with translational potential will lead to improved treatment for stress-related sleep disturbances.

Multivariable Model to Predict an ACTH Stimulation Test to Diagnose Adrenal Insufficiency Using Previous Test Results.

Context: The adrenocorticotropin hormone stimulation test (AST) is used to diagnose adrenal insufficiency, and is often repeated in patients when monitoring recovery of the hypothalamo-pituitary-adrenal axis. Objective: To develop and validate a prediction model that uses previous AST results with new baseline cortisol to predict the result of a new AST. Methods: This was a retrospective, longitudinal cohort study in patients who had undergone at least 2 ASTs, using polynomial regression with backwards variable selection, at a Tertiary UK adult endocrinology center. Model was developed from 258 paired ASTs over 5 years in 175 adults (mean age 52.4 years, SD 16.4), then validated on data from 111 patients over 1 year (51.8, 17.5) from the same center, data collected after model development. Candidate prediction variables included previous test baseline adrenocorticotropin hormone (ACTH), previous test baseline and 30-minute cortisol, days between tests, and new baseline ACTH and cortisol used with calculated cortisol/ACTH ratios to assess 8 candidate predictors. The main outcome measure was a new test cortisol measured 30 minutes after Synacthen administration. Results: Using 258 sequential ASTs from 175 patients for model development and 111 patient tests for model validation, previous baseline cortisol, previous 30-minute cortisol and new baseline cortisol were superior at predicting new 30-minute cortisol (R2 = 0.71 [0.49-0.93], area under the curve [AUC] = 0.97 [0.94-1.0]) than new baseline cortisol alone (R2 = 0.53 [0.22-0.84], AUC = 0.88 [0.81-0.95]). Conclusion: Results of a previous AST can be objectively combined with new early-morning cortisol to predict the results of a new AST better than new early-morning cortisol alone. An online calculator is available at https://endocrinology.shinyapps.io/sheffield_sst_calculator/ for external validation.

Symptomatic benefits of testosterone treatment in patient subgroups: a systematic review, individual participant data meta-analysis, and aggregate data meta-analysis.

BACKGROUND: Testosterone replacement therapy is known to improve sexual function in men younger than 40 years with pathological hypogonadism. However, the extent to which testosterone alleviates sexual dysfunction in older men and men with obesity is unclear, despite the fact that testosterone is being increasingly prescribed to these patient populations. We aimed to evaluate whether subgroups of men with low testosterone derive any symptomatic benefit from testosterone treatment. METHODS: We did a systematic review and meta-analysis to evaluate characteristics associated with symptomatic benefit of testosterone treatment versus placebo in men aged 18 years and older with a baseline serum total testosterone concentration of less than 12 nmol/L. We searched major electronic databases (MEDLINE, Embase, Science Citation Index, and the Cochrane Central Register of Controlled Trials) and clinical trial registries for reports published in English between Jan 1, 1992, and Aug 27, 2018. Anonymised individual participant data were requested from the investigators of all identified trials. Primary (cardiovascular) outcomes from this analysis have been published previously. In this report, we present the secondary outcomes of sexual function, quality of life, and psychological outcomes at 12 months. We did a one-stage individual participant data meta-analysis with a random-effects linear regression model, and a two-stage meta-analysis integrating individual participant data with aggregated data from studies that did not provide individual participant data. This study is registered with PROSPERO, CRD42018111005. FINDINGS: 9871 citations were identified through database searches. After exclusion of duplicates and publications not meeting inclusion criteria, 225 full texts were assessed for inclusion, of which 109 publications reporting 35 primary studies (with a total 5601 participants) were included. Of these, 17 trials provided individual participant data (3431 participants; median age 67 years [IQR 60-72]; 3281 [97%] of 3380 aged ≥40 years) Compared with placebo, testosterone treatment increased 15-item International Index of Erectile Function (IIEF-15) total score (mean difference 5·52 [95% CI 3·95-7·10]; τ2=1·17; n=1412) and IIEF-15 erectile function subscore (2·14 [1·40-2·89]; τ2=0·64; n=1436), reaching the minimal clinically important difference for mild erectile dysfunction. These effects were not found to be dependent on participant age, obesity, presence of diabetes, or baseline serum total testosterone. However, absolute IIEF-15 scores reached during testosterone treatment were subject to thresholds in patient age and baseline serum total testosterone. Testosterone significantly improved Aging Males' Symptoms score, and some 12-item or 36-item Short Form Survey quality of life subscores compared with placebo, but it did not significantly improve psychological symptoms (measured by Beck Depression Inventory). INTERPRETATION: In men aged 40 years or older with baseline serum testosterone of less than 12 nmol/L, short-to-medium-term testosterone treatment could provide clinically meaningful treatment for mild erectile dysfunction, irrespective of patient age, obesity, or degree of low testosterone. However, due to more severe baseline symptoms, the absolute level of sexual function reached during testosterone treatment might be lower in older men and men with obesity. FUNDING: National Institute for Health and Care Research Health Technology Assessment Programme.

Using Generative Artificial Intelligence Tools in Cosmetic Surgery: A Study on Rhinoplasty, Facelifts, and Blepharoplasty Procedures.

Artificial intelligence (AI), notably Generative Adversarial Networks, has the potential to transform medical and patient education. Leveraging GANs in medical fields, especially cosmetic surgery, provides a plethora of benefits, including upholding patient confidentiality, ensuring broad exposure to diverse patient scenarios, and democratizing medical education. This study investigated the capacity of AI models, DALL-E 2, Midjourney, and Blue Willow, to generate realistic images pertinent to cosmetic surgery. We combined the generative powers of ChatGPT-4 and Google's BARD with these GANs to produce images of various noses, faces, and eyelids. Four board-certified plastic surgeons evaluated the generated images, eliminating the need for real patient photographs. Notably, generated images predominantly showcased female faces with lighter skin tones, lacking representation of males, older women, and those with a body mass index above 20. The integration of AI in cosmetic surgery offers enhanced patient education and training but demands careful and ethical incorporation to ensure comprehensive representation and uphold medical standards.

Comparing the Efficacy of Large Language Models ChatGPT, BARD, and Bing AI in Providing Information on Rhinoplasty: An Observational Study.

Background: Large language models (LLMs) are emerging artificial intelligence (AI) technologies refining research and healthcare. However, the impact of these models on presurgical planning and education remains under-explored. Objectives: This study aims to assess 3 prominent LLMs-Google's AI BARD (Mountain View, CA), Bing AI (Microsoft, Redmond, WA), and ChatGPT-3.5 (Open AI, San Francisco, CA) in providing safe medical information for rhinoplasty. Methods: Six questions regarding rhinoplasty were prompted to ChatGPT, BARD, and Bing AI. A Likert scale was used to evaluate these responses by a panel of Specialist Plastic and Reconstructive Surgeons with extensive experience in rhinoplasty. To measure reliability, the Flesch Reading Ease Score, the Flesch-Kincaid Grade Level, and the Coleman-Liau Index were used. The modified DISCERN score was chosen as the criterion for assessing suitability and reliability. A t test was performed to calculate the difference between the LLMs, and a double-sided P-value <.05 was considered statistically significant. Results: In terms of reliability, BARD and ChatGPT demonstrated a significantly (P < .05) greater Flesch Reading Ease Score of 47.47 (±15.32) and 37.68 (±12.96), Flesch-Kincaid Grade Level of 9.7 (±3.12) and 10.15 (±1.84), and a Coleman-Liau Index of 10.83 (±2.14) and 12.17 (±1.17) than Bing AI. In terms of suitability, BARD (46.3 ± 2.8) demonstrated a significantly greater DISCERN score than ChatGPT and Bing AI. In terms of Likert score, ChatGPT and BARD demonstrated similar scores and were greater than Bing AI. Conclusions: BARD delivered the most succinct and comprehensible information, followed by ChatGPT and Bing AI. Although these models demonstrate potential, challenges regarding their depth and specificity remain. Therefore, future research should aim to augment LLM performance through the integration of specialized databases and expert knowledge, while also refining their algorithms.

Real world evidence supports waking salivary cortisone as a screening test for adrenal insufficiency.

OBJECTIVE: Worldwide, adults and children are at risk of adrenal insufficiency largely due to infectious diseases and adrenal suppression from use of anti-inflammatory glucocorticoids. Home waking salivary cortisone is an accurate screening test for adrenal insufficiency, it has potential to reduce costs, and patients prefer it to the adrenocorticotropin (ACTH) (synacthen) stimulation test. We carried out a service evaluation of home waking salivary cortisone in clinical care to identify implementation barriers. DESIGN, PATIENTS AND MEASUREMENTS: Service evaluation in a centre where 212 patients referred for adrenal insufficiency had a waking salivary cortisone. Problems encountered during testing were recorded and patient feedback, via focus groups, collected. RESULTS: From all patients providing a waking salivary cortisone 55% had a normal test, 23% adrenal suppression, and 22% an equivocal result requiring a clinical centre ACTH stimulation test. The median (interquartile range [IQR]) for the time of the saliva sample was 07:40 (07:00-08:40). The median (IQR) days between collection and (i) delivery to local laboratory was 1 (0.25-2) day; (ii) reporting by local laboratory was 13 (11-18) days. Patients considered the test is "easy to do" and preferred it to the inpatient ACTH stimulation test. The principal challenge to clinical implementation was results reporting to clinicians due to delays at the local laboratory. CONCLUSIONS: This service evaluation provides real-world evidence that home waking salivary cortisone is an effective, practical screening test for adrenal insufficiency. It identified key barriers to testing implementation that need to be addressed when introducing the test to a health service.

Developing oral chronotherapy for cortisol replacement in congenital adrenal hyperplasia.

The sun imposes a 24-h periodicity to life and circadian rhythms have evolved to maintain homoeostasis through the day/night cycle. In humans, there is a central clock that controls the sleep/wake cycle which is paralleled metabolically by a fast/feed cycle. The clock maintains homoeostasis by synchronising metabolism to the time of feeding. Loss of synchrony between the clock and hormonal rhythms results in loss of homoeostasis as evidenced by obesity, depression, and diabetes in people undertaking shift work. Cortisol has a distinct circadian rhythm; peaking on waking and low at sleep onset. Loss of this rhythm in adrenal insufficiency is associated with a poor quality of life and increased mortality. To replace the cortisol rhythm requires chronotherapy and for this you need to define the key parameters of the target rhythm, create a formulation to replicate that rhythm, and then prove clinical benefit. The physiology of hormones is more complex than that of nonnative drugs. Hormones are secreted with varied rhythms, bound to multiple cognate binding proteins, and actively transported and cleared through enzymatic pathways in multiple organs. We have examined the diurnal rhythm of cortisol in healthy volunteers, created physiologically-based pharmacokinetic models, and tested various oral delayed and sustained formulations of hydrocortisone (development name, Chronocort) in clinical trials. The outcome from this work was the manufacture of modified-release hydrocortisone hard capsules (tradename Efmody, Diurnal Ltd), that replicate the cortisol diurnal rhythm and improve the disease control of congenital adrenal hyperplasia the commonest hereditary form of adrenal insufficiency.

Measurement of mouse head and neck tumors by automated analysis of CBCT images.

Animal experiments are often used to determine effects of drugs and other biological conditions on cancer progression, but poor accuracy and reproducibility of established tumor measurement methods make results unreliable. In orthotopic mouse models of head and neck cancer, tumor volumes approximated from caliper measurements are conventionally used to compare groups, but geometrical challenges make the procedure imprecise. To address this, we developed software to better measure these tumors by automated analysis of cone-beam computed tomography (CBCT) scans. This allows for analyses of tumor shape and growth dynamics that would otherwise be too inaccurate to provide biological insight. Monitoring tumor growth by calipers and imaging in parallel, we find that caliper measurements of small tumors are weakly correlated with actual tumor volume and highly susceptible to experimenter bias. The method presented provides a unique window to sources of error in a foundational aspect of preclinical head and neck cancer research and a valuable tool to mitigate them.

Evaluating Chatbot Efficacy for Answering Frequently Asked Questions in Plastic Surgery: A ChatGPT Case Study Focused on Breast Augmentation.

BACKGROUND: The integration of artificial intelligence (AI) and machine learning (ML) technologies into healthcare is transforming patient-practitioner interaction and could offer an additional platform for patient education and support. OBJECTIVES: This study investigated whether ChatGPT-4 could provide safe and up-to-date medical information about breast augmentation that is comparable to other patient information sources. METHODS: ChatGPT-4 was asked to generate 6 commonly asked questions regarding breast augmentation and respond to them. Its responses were qualitatively evaluated by a panel of specialist plastic and reconstructive surgeons and reconciled with a literature search of 2 large medical databases for accuracy, informativeness, and accessibility. RESULTS: ChatGPT-4 provided well-structured, grammatically accurate, and comprehensive responses to the questions posed; however, it was limited in providing personalized advice and sometimes generated inappropriate or outdated references. ChatGPT consistently encouraged engagement with a specialist for specific information. CONCLUSIONS: Although ChatGPT-4 showed promise as an adjunct tool in patient education regarding breast augmentation, there are areas requiring improvement. Additional advancements and software engineering are needed to enhance the reliability and applicability of AI-driven chatbots in patient education and support systems.