Background: Large language models (LLMs) have demonstrated potential in medical diagnostics, but their accuracy in complex cases remains a subject of investigation. DeepSeek-R1, an open-source model with advanced reasoning capabilities, has gained global attention. This study evaluates the diagnostic performance of DeepSeek-R1 compared to GPT-4 in complex clinical cases. Materials and methods: A historical control study was conducted using 100 clinicopathologic cases from the New England Journal of Medicine (NEJM), published between 18 August 2022, and 30 January 2025. Each case was processed using DeepSeek-R1 with a structured diagnostic prompt. The model’s performance was assessed based on final diagnosis accuracy, differential diagnosis inclusion rate, ranking of correct diagnoses, and differential quality scores. Results were statistically compared to previously published GPT-4 performance data using chi-square, Mann-Whitney U, and t-tests. Results: DeepSeek-R1 correctly matched the final diagnosis in 35% of cases (35/100), which was comparable to GPT-4’s accuracy (39%; P = 0.634). However, DeepSeek-R1 included the correct diagnosis in its differential list in 48% of cases, significantly lower than GPT-4 (64%; P = 0.036). DeepSeek-R1 generated longer differential diagnoses (11.9 ± 2.0 vs. 9.0 ± 1.4; P = 0.000004) but maintained a similar mean rank for correct diagnoses (1.8 ± 2.2 vs. 2.5 ± 2.5; P = 0.288566) and equivalent differential quality scores (4.2 ± 0.10 vs. 4.2 ± 1.3; P = 0.099667). Conclusion: DeepSeek-R1 exhibits diagnostic accuracy comparable to GPT-4 while generating more diverse differential diagnoses. Its open-source nature and innovative reasoning strategies may enhance medical AI applications. Future studies should explore real-world clinical integration and refinement of differential diagnosis prioritization.

Poor sleep quality is associated with a decrease in quality of life in patients with major burn scars, combined with pruritus and pain. Few interventions have been reported to improve the sleep quality of patients with scars. In the current prospective cohort study, we investigated the efficacy of CO₂-ablative fractional laser (AFL) surgery vs conventional surgery in post-burn patients with hypertrophic scars with sleep quality as the primary study outcome. In total 68 consecutive patients undergoing scar surgical treatment were recruited, including a CO₂-AFL surgery cohort (n = 35) and a conventional surgery cohort (n = 33). A subgroup from the AFL cohort was selected. Sleep quality, pain and pruritus were evaluated. Multiple linear regression analyses were performed to reveal the effect of CO₂-AFL surgery. The CO₂-AFL surgery cohort had significantly lower Pittsburgh sleep quality index (PSQI) global scores than the conventional surgery cohort after the last surgical treatment. In the subgroup of patients receiving hardware sleep monitoring, CO₂-AFL markedly increased deep sleep time, deep sleep efficiency and reduced initial sleep latency. Compared to the conventional surgery cohort, the CO₂-AFL cohort presented significantly lower pain and pruritus scores. Correlation analysis showed pain and pruritus were significantly associated with PSQI scores, and there were also significant correlations between pain and pruritus scores. Multiple linear regression analysis showed that surgery method was negatively linearly correlated with visual analog scale (VAS) pain score, brief pain inventory (BPI) total, VAS pruritus score, 5-D itch scale total, four-item itch questionnaire (FIIQ) total and PSQI total. CO₂-AFL surgery significantly improved sleep quality and reduced pain and pruritus of hypertrophic scar patients. The alleviation of sleep disorder was associated with improvement of deep sleep quality including deep sleep time and deep sleep deficiency. The Chinese Clinical Trial Registry (ChiCTR200035268) approved retrospectively registration on 5 Aug 2020.

Hypertrophic scars cause impaired skin appearance and function, seriously affecting physical and mental health. Due to medical ethics and clinical accessibility, the collection of human scar specimens is frequently restricted, and the establishment of scar experimental animal models for scientific research is urgently needed. The four most commonly used animal models of hypertrophic scars have the following drawbacks: the rabbit ear model takes a long time to construct; the immunodeficient mouse hypertrophic scar model necessitates careful feeding and experimental operations; female Duroc pigs are expensive to purchase and maintain, and their large size makes it difficult to produce a significant number of models; and mouse scar models that rely on tension require special skin stretch devices, which are often damaged and shed, resulting in unstable model establishment. Our group overcame the shortcomings of previous scar animal models and created a new mouse model of hypertrophic scarring induced by suture anchoring at the wound edge. We utilized suture anchoring of incisional wounds to impose directional tension throughout the healing process, restrain wound contraction, and generate granulation tissue, thus inducing scar formation. Dorsal paired incisions were generated in mice, with wound edges on the upper back sutured to the rib cage and the wound edges on the lower back relaxed as a control. Macroscopic manifestation, microscopic histological analysis, mRNA sequencing, bioinformatics, and in vitro cell assays were also conducted to verify the reliability of this method. Compared with those in relaxed controls, the fibrotic changes in stretched wounds were more profound. Histologically, the stretched scars were hypercellular, hypervascular, and hyperproliferative with disorganized extracellular matrix deposition, and displayed molecular hallmarks of hypertrophic fibrosis. In addition, the stretched scars exhibited transcriptional overlap with mechanically stretched scars, and human hypertrophic and keloid scars. Phosphatidylinositol 3-kinase-serine/threonine-protein kinase B signaling was implicated as a profibrotic mediator of apoptosis resistance under suture-induced tension. This straightforward murine model successfully induces cardinal molecular and histological features of pathological hypertrophic scarring through localized suture tension to inhibit wound contraction. The model enables us to interrogate the mechanisms of tension-induced fibrosis and evaluate anti-scarring therapies.

Quiescent hair follicle stem cells (HFSCs) reside in specialized bulge niche where they undergo activation and differentiation upon sensing niche-dependent signals during hair follicle (HF) homeostasis and wound repair. The underlying mechanism of HFSCs and bulge niche maintenance is poorly understood. Our previous study has reported that a transcription factor, forkhead box P1 (Foxp1), functions to maintain the quiescence of HFSCs. Here, we further discovered that forkhead box P4 (Foxp4), a close family member of Foxp1, had similar expression profiles in various components of HFs and formed a complex with Foxp1 in vitro and in vivo. The HF-specific deficiency of Foxp4 resulted in the precocious activation of HFSCs during hair cycles. In contrast to single Foxp1 or Foxp4 conditional knockout (cKO) mice, Foxp1/4 double cKO exerted an additive effect in the spectrum and severity of phenotypes in HFSC activation, hair cycling acceleration and hair loss, coupled with remarkable downregulation of fibroblast growth factor 18 (Fgf18) and bone morphogenetic protein 6 (Bmp6) expression in bulge cells. In addition, the double KO of Foxp1/4 induced the apoptosis of K6-positive (K6+) inner bulge cells, a well-established stem cell (SC) niche, thus resulting in the destruction of the bulge SC niche and recurrent hair loss. Our investigation reveals the synergistic role of Foxp1/4 in sustaining K6+ niche cells for the quiescence of HFSCs.