“Acne scars” is a category, not a diagnosis. The classification physicians use — icepick, boxcar, rolling, plus raised hypertrophic scars — exists because each shape reflects a different injury below the surface and answers to a different tool. Most scarred faces carry a mixture, which is why one-device treatment plans disappoint.
WhatsApp Dr Sin Yong →The classification literature — from Jacob's original system to its modern updates — is built on one clinical fact: scar morphology encodes the injury beneath [1,2]. An icepick scar is a deep epithelial tract punched by a ruptured follicle; a boxcar is a block of lost dermis with intact edges; a rolling scar is skin of near-normal quality yanked downward by fibrous tethers to deeper tissue. Same disease, three different structural problems — and a treatment brilliant for one is irrelevant for another. Resurface an icepick and you polish the rim of a well. Subcise a boxcar and nothing changes, because nothing was tethered.
The evidence-based management literature maps the toolkit [2,3]: icepick scars respond to focal chemical reconstruction (TCA CROSS) and punch techniques; boxcars to fractional ablative resurfacing; rolling scars to subcision that releases the tethers, often with fillers or biostimulation supporting the released skin; hypertrophic and keloid scars — a different biology entirely — to the injection-based pathway on the keloid page. Energy-based collagen remodelling — fractional CO2, RF microneedling — runs across types as the field-improvement layer. This is why Dr Sin Yong's Tetra Pro SCAR3 programme and 4D scar reconstruction begin with a mapped scar census, not a device booking — the plan is a sequence assembled per face. Flat dark marks are a different problem: see PIH.
First: control the acne before reconstructing its damage — active disease keeps manufacturing new scars behind every repair, so the acne programme comes first when lesions are still appearing. Second: expect a campaign, not an event. The literature is consistent that meaningful atrophic scar revision is staged — collagen remodelling matures over months between sessions — and that improvement, not erasure, is the honest endpoint [2,3]. Anyone promising scar removal in one session is describing neither the evidence nor the biology.
Scar creams and oils on established atrophic scars — a structural deficit does not refill from the surface. Aggressive scrubbing or home dermarollers — uncontrolled micro-injury on scar-prone skin risks new damage, including PIH in Asian skin. One-device-fixes-all packages — a face carrying icepick, boxcar and rolling scars needs a sequenced mix, and a clinic that offers only one modality will treat every scar as if it were the type their machine handles. And waiting for scars to fade — atrophic scars are permanent without intervention; time softens edges, not deficits.
“A treatment brilliant for one scar type is irrelevant for another — resurfacing an icepick scar just gives the well a polished rim.”
— Dr Sin Yong
Stretch test and light: pull the skin taut — rolling scars flatten dramatically, boxcars soften slightly, icepicks barely change. Overhead light exaggerates rolling scars. A clinical mapping confirms and counts them.
No modality erases scars to unmarked skin — the honest, evidence-based endpoint is substantial, staged improvement. Distrust the word 'removal' in scar marketing.
A needle-based release of the fibrous bands tethering rolling scars — freeing the skin so it can sit level again. It targets the mechanism, which is why it beats surface treatment for that type.
Yes, with fractional settings chosen for pigment-prone skin and strict aftercare — post-treatment pigmentation is the known risk, managed by conservative parameters and photoprotection.
Control first, reconstruct second. Active acne keeps creating new scars behind each repair — sequencing protects both the result and the spend.
Side and overhead light casts shadows into depressions — rolling scars especially. That is also the honest way to photograph progress: same light, same angle, every time.