The term "acute diffuse and total alopecia" (ADTA) has been often used as a synonym for self-regressing severe alopecia areata (AA). However, ADTA is originally defined as a rapidly-progressive subtype of AA (RP-AA) with short recovery time and favorable prognosis irrespective of interventions. Indeed, a subpopulation of ADTA recovers spontaneously. We focused on this unique subset of AA, which we coined as "self-healing ADTA" (sADTA). Prompt and accurate differentiation of sADTA from other RP-AA is important to avoid unnecessary treatments, which is still challenging due to the lack of predictive diagnostic hallmarks. In this study, 18 sADTA patients were retrospectively analyzed to delineate their demographics and clinical features, including gentle hair pull test and trichoscopic findings, followed by statistical comparison with those of RP-AA. All sADTA cases were female and the average age was 38.1 AE 15.9 years. The progression of hair loss areas peaked at 3.6 AE 1.5 months, and complete hair regrowth was achieved in 7.9 AE 1.7 months. The identified factors supporting the diagnosis of sADTA included being female; the absence of scalp pain and itch; the absence of extra-scalp hair loss; club hair predominance in hair pull test; predominant short vellus hairs; and increase in vacant follicular ostia on trichoscopy. Subsequently, a scoring system for early diagnosis of sADTA was developed by counting the number of six identified factors present in a tested case. When analyzed by the developed system, all sADTA cases, including an additional four cases, had scores of 4 or above, while all RP-AA cases had scores below 3 except one case. Therefore, the system successfully differentiated sADTA from RP-AA (P < 0.01). Despite some technical limitations, the current study suggested that sADTA is a distinctive entity with unique pathophysiology and that early diagnosis before intervention is feasible based on the characteristics.