Patient Perspectives regarding Artificial Intelligence-Driven Trichoscopy for Alopecia Evaluation Authors: Caitlin A. Kearney, Anna L. Brinks, Carli N. Lawrence, Jerry Shapiro, Mario E. Lacouture, Kristen I. Lo Sicco Published: 2025 ABSTRACT AbstractIntroductionAlopecia treatment can improve patients’ quality of life, but slow hair regrowth makes progress difficult to detect. Trichoscopic measurements (TMs; hair density, hairs/cm2; hair caliber, µm) provide an objective method for monitoring treatment response. Artificial intelligence (AI) enables trichoscopic analysis without hair clippings. This study examines the perspectives of alopecia patients regarding AI-driven trichoscopy.MethodsWe conducted an IRB-approved survey study and chart review of 101 patients at the NYU Langone hair clinic undergoing AI-driven trichoscopy. Pre- and post-visit surveys measured self-perceived alopecia severity and quality-of-life impact before and after reviewing TMs with their dermatologist. Opinions regarding AI-driven trichoscopy were assessed via Likert scales. Demographics, medical history, and clinician-assessed severity were extracted from medical charts.ResultsMost participants were female (74.3%), white (73.3%), and had nonscarring alopecia (89%). After reviewing TMs, 33.7% of participants reported feeling less bothered by their alopecia, 39.6% perceived reduced severity, and 54.5% noted increased treatment motivation. Among follow-up patients, concordance between patient and physician assessments of interval change in hair loss improved after viewing TMs (p = 0.0075). Patients endorsed enhanced alopecia care and greater confidence in treatment recommendations with the use of AI-driven trichoscopy.ConclusionPatients with alopecia generally view AI-driven trichoscopy favorably, noting improved quality of care and greater motivation to continue therapy. Introduction Successful alopecia treatment can improve quality of life, although slow regrowth makes progress detection difficult [1, 2]. Scalp photography has been shown to reduce alopecia-related anxiety, increase treatment motivation, and improve congruence between clinician and patient assessments [1]. However, photographic quantification of alopecia presents challenges due to variations in hair and scalp appearance, photographic conditions, and perceptual limitations [1, 3]. Trichoscopic measurements (TMs), including hair density (hairs/cm2) and caliber (µm), provide an objective method for detecting treatment response [3]. Artificial intelligence (AI) enables efficient trichoscopic analysis without the need for hair clipping, unlike traditional phototrichograms [3, 4]. This study examines patients’ perspectives regarding AI-driven trichoscopy and its impact on self-perception of alopecia. Methods We conducted an Institutional Review Board-approved mixed-methods survey study and chart review of patients with alopecia undergoing AI-driven trichoscopy (Canfield HairMetrix®) at the NYU Langone Health hair clinic (online suppl. Fig. 1; for all online suppl. material, see https://doi.org/10.1159/000550043). Participants completed a pre-survey before trichoscopy and a post-survey after discussing trichoscopy results with their dermatologist. Both surveys assessed self-perceived alopecia severity and quality of life impact. The post-survey also evaluated opinions regarding AI-driven trichoscopy via 5-point Likert scales. Demographics, medical history, clinician-assessed severity, and TMs were extracted from medical charts. For follow-up patients with prior TMs, self-assessments of alopecia interval change since the previous visit were recorded before and after viewing TMs with their dermatologist. These self-assessments were compared with the clinician’s assessment of the interval change, as documented in the medical record. Concordances between patient and clinician assessments of interval change were evaluated using Cohen’s kappa, and the change in concordance after viewing TMs was tested for significance (p < 0.05) via bootstrapping with 2,000 replicates [5]. Statistical analyses were performed in RStudio (version 2025.05.0+496). Results Of the 103 patients approached, 101 participated (98.1%). The majority of participants were female (74.3%) and white (73.3%) (Table 1). The most common alopecia diagnoses were androgenetic alopecia (AGA) (87.1%) and telogen effluvium (48.5%). After reviewing TMs with their dermatologist, 33.7% of participants reported feeling less bothered by alopecia, 39.6% reported decreased perceived alopecia severity, and 54.5% reported increased treatment motivation (Table 1). Among follow-up patients, the agreement between physician assessment and participant self-assessment of alopecia progression since the last visit improved after participants viewed their TMs, increasing from fair (κ = 0.217, p = 0.064) to moderate (κ = 0.437, p = 0.0002), a significant change (p = 0.0075; Table 1) [5]. Most patients viewed TMs positively, endorsing enhanced care and greater confidence in treatment recommendations (Fig. 1). Average opinions about AI-driven trichoscopy were similar between participants with scarring alopecias (n = 11) and those with nonscarring alopecias (n = 90; online suppl. Table 1). Table 1.Participant demographics and responses to viewing AI-driven TMsParticipant demographics and responses to viewing TMs (N = 101)Age, mean±SD, years47.1±18.0Participant-reported age at alopecia onset37.4±18.4Race and ethnicity, n (%) Asian8 (7.9) Hispanic, Latinx, or Spanish origin1 (1.0) Middle Eastern or North African2 (2.0) White74 (73.3) Unknown17 (16.8)Sex, n (%) Male26 (25.7) Female75 (74.3)Alopecia diagnoses, n (%)a AGA88 (87.1) Telogen effluvium49 (48.5) Alopecia areata5 (5.0) Frontal fibrosing alopecia6 (5.9) Lichen planopilaris5 (5.0) Discoid lupus erythematosus1 (1.0) Endocrine therapy-induced alopecia1 (1.0) Radiation-induced alopecia1 (1.0) Fibrosing alopecia in a pattern distribution3 (3.0) Seborrheic dermatitis1 (1.0) Trichotillomania2 (2.0)Current alopecia treatments, n (%) Topical medications77 (76.2) Oral medications80 (79.2) Adjunctive treatments (e.g., PRP, LLLT)24 (23.8) None11 (10.9)Change in patient attitude toward alopecia after viewing hair counts, n (%) Less bothered by alopecia34 (33.7) No change60 (59.4) More bothered by alopecia7 (6.9)Change in patient treatment motivation after viewing hair counts, n (%) More motivated to treat alopecia55 (54.5) No change46 (45.5)Change in self-perceived alopecia severity after viewing hair counts, n (%) Less severe40 (39.6) The same53 (52.5) More severe8 (7.9)Concordance between patient and physician assessment of alopecia progression (N = 66)bKappa statistic p valueConcordance before patients viewed TMs0.2170.064Concordance after patients viewed TMs0.4370.0002*Bootstrap for concordance improvement​0.0075*PRP, platelet-rich plasma; LLLT, low-level light therapy. aPercentages do not add up to 100% as participants may have more than one type of alopecia. bFor follow-up patients with prior TMs, self-assessments of alopecia interval change since the previous visit (worse, the same, better) were recorded before and after viewing TMs with their dermatologist. These self-assessments were compared with the clinician’s assessment of interval change (worsening, stable, improving) as documented in the medical record for the survey visit. Concordances between patient and clinician assessments were evaluated using Cohen’s kappa, and the change in concordance after viewing TMs was tested for significance (p < 0.05) via bootstrapping with 2,000 replicates. Participant demographics and responses to viewing AI-driven TMs PRP, platelet-rich plasma; LLLT, low-level light therapy. aPercentages do not add up to 100% as participants may have more than one type of alopecia. bFor follow-up patients with prior TMs, self-assessments of alopecia interval change since the previous visit (worse, the same, better) were recorded before and after viewing TMs with their dermatologist. These self-assessments were compared with the clinician’s assessment of interval change (worsening, stable, improving) as documented in the medical record for the survey visit. Concordances between patient and clinician assessments were evaluated using Cohen’s kappa, and the change in concordance after viewing TMs was tested for significance (p < 0.05) via bootstrapping with 2,000 replicates. Heat map of participant agreement with statements regarding the use of AI-driven trichoscopy in their monitoring of alopecia (N = 101). Participants indicated their agreement with opinion statements using a 5-point Likert scale. Hair counts refer to TMs produced by the AI analysis, including terminal hair density and caliber. Greater color intensity corresponds to a higher frequency of response, as indicated by the percentage of participants notated within each square. *Questions only completed by patients with prior TMs, enabling discussion of interval change (N = 51). Among 51 participants with prior TMs, 23.5% were bothered by stable counts, 60.8% by decreased counts, and 11.7% reported thinking about their TM numbers often. Of the 12 participants bothered by stable TMs, 4 (33.3%) were on stable alopecia medication regimens for ≥6 months (mean 15.1 months; standard deviation 6.0 months). Participants with scarring alopecias (n = 3) reported less bother with stable counts than those with nonscarring alopecias (n = 48; average Likert score 1.67 and 3.06, respectively; online suppl. Table 1). Discussion Our findings suggest that patients perceive AI-driven trichoscopy as a valuable tool for clinical evaluation of alopecia [2]. Trichoscopic data may enhance patients’ sense of control and reduce anxiety; however, decreasing TMs could exacerbate distress. Most participants with prior TMs reported bother with decreased counts (60.8%). Discussion of negative results with patients could increase visit duration, prompt additional calls to providers, and heighten dissatisfaction with treatment. Reassuringly, few patients (11.7%) reported rumination on their TMs between visits. Nine of the 12 participants (75%) bothered by stable counts had AGA, highlighting the need to counsel on stability as a treatment goal in AGA. In addition, the observed difference in bother regarding stable TMs between participants with scarring alopecias and nonscarring alopecias may reflect differing treatment expectations. Limitations include a small sample size, convenience sampling, the use of a non-validated survey instrument, self-report bias, and variations in health literacy. Participants were primarily white, female, and had nonscarring alopecias, limiting generalizability beyond these demographics and to scarring alopecias. This study demonstrates patient acceptance of AI-driven trichoscopy as an alopecia monitoring tool. AI-driven trichoscopy provides an efficient, noninvasive method to capture detailed, objective data on alopecia and its changes over time. However, several barriers may limit its integration into clinical practice, including equipment costs and the additional time or personnel needed to perform measurements. The potential for expenses to be passed on to patients could further restrict its accessibility and routine use. Future studies should compare AI-driven trichoscopy with standard trichoscopy and global scalp photography to assess whether these methods similarly influence alopecia-related anxiety, treatment motivation, and the alignment between clinician and patient assessments of alopecia progression. Additional studies involving patients with scarring alopecias and more diverse populations are needed to better capture the potential range of perspectives on this emerging technology. Statement of Ethics This study protocol was reviewed and approved by the NYU Langone Health Institutional Review Board (s24-01747). The authors attest that written informed consent was obtained from all patients participating in this study. Conflict of Interest Statement Dr. Lacouture is an investigator for Hoth Therapeutics, a consultant for Hoth Therapeutics and Marengo Therapeutics, and the president of the Oncodermatology Society. Dr. Lo Sicco is an investigator for Pfizer, a consultant for Pfizer, Veradermics, and Acquis, and a board member of the Scarring Alopecia Foundation and American Hair Research Society. Dr. Lo Sicco was a member of the journal’s Editorial Board at the time of submission. Dr. Shapiro is an investigator for Pfizer and a consultant for Pfizer, Lilly, Replicel Life Science, Regen Lab, Thirty Madison, DS Laboratories, and Almirall. Authors Kearney, Brinks, and Lawrence have no conflicts of interest to declare. Funding Sources This study was not supported by any sponsor or funder. Author Contributions Caitlin A. Kearney: conceptualization, investigation, methodology, formal analysis, and writing – original draft, review, and editing; Anna L. Brinks and Carli N. Lawrence: investigation and writing – original draft, review, and editing; Jerry Shapiro: project administration and writing – review and editing; Mario E. Lacouture and Kristen I. Lo Sicco: conceptualization, project administration, supervision, and writing – review and editing.