Acne scars: the different types, and why the same treatment does not work on all of them

When a spot heals, the skin has been through an inflammatory event. Two separate things can be left behind, and they are produced by different mechanisms in different layers.

The first is a pigment or vascular change. The skin's pigment-producing cells respond to inflammation by depositing extra melanin, leaving a brown, grey-brown or slate-coloured mark. Separately, the small blood vessels that dilated during the inflammation can stay dilated and visible for a while, leaving a pink or red mark. Neither of these involves any change to the shape of the skin. Run a fingertip over them and the surface is flat. In dermatological language these are post-inflammatory hyperpigmentation and post-inflammatory erythema — the first is a pigment problem, the second is a blood-vessel problem, and calling either of them a scar is technically wrong even though everybody does it.

The second is an actual scar: the skin's repair process ended with either less collagen than it started with, leaving a depression, or more than it started with, leaving a raised lump. The surface geometry has changed. You can feel it with your eyes closed. This is what the word scar is supposed to mean, and it is the category that does not resolve by waiting.

Almost every bad decision in this area comes from mistaking the first for the second. People buy resurfacing treatments for marks that were going to fade anyway, and then credit the treatment. Or they wait patiently for something to fade that structurally cannot, and lose years. Getting the category right first is worth more than any product choice that follows it.

Melanocytes sit at the base of the epidermis and produce melanin. Inflammation is one of the signals that switches them into overdrive, which is the same basic mechanism behind a suntan — the skin responds to an insult by making pigment. After a spot resolves, that extra pigment can persist in the epidermis, and in some cases pigment drops into the dermis below, where it sits deeper and is much slower to clear.

This matters because depth predicts timeline. Pigment in the epidermis turns over with the epidermis. Pigment that has fallen into the dermis has no such conveyor belt and can linger far longer. That is one reason two people with visually similar marks can have very different experiences of waiting.

It is well recognised in dermatology that post-inflammatory hyperpigmentation shows up more readily and persists longer in deeper skin tones, simply because there is more melanin-making machinery to respond. This is not a cosmetic footnote — it changes which treatments are appropriate, because several resurfacing approaches carry their own risk of provoking further pigment change in exactly the skin types most prone to it. It is the clearest example of why generic internet advice fails here, and why the assessment has to be individual.

Two other things are consistently emphasised for pigment marks. Ultraviolet exposure re-stimulates the same pigment cells, so an unprotected mark is being topped up in the background while you wait for it to fade. And the active acne itself is the source: every new inflamed lesion in the same area deposits a fresh mark, so a pigmentation plan on top of uncontrolled breakouts is a treadmill.

The red or purple-pink marks are a different problem with a different cause. During inflammation the small vessels in the dermis dilate and can be damaged. Afterwards, the visible redness is blood in dilated or newly formed capillaries close to the surface — not pigment at all. It tends to be more conspicuous in lighter skin, where there is less overlying melanin to mask it.

The practical consequence is that treatments aimed at pigment do very little for it, because there is no excess pigment to address. Vascular marks are typically approached through a completely separate family of options, and a clinician distinguishes them precisely because the mismatch wastes time.

There is a crude at-home discriminator that dermatologists use in a more controlled form. Press gently on the mark — with a clean fingertip, or the edge of a glass slide, which is the clinical version. A vascular mark blanches: the pressure pushes blood out of the vessels and the colour momentarily vanishes, returning when you release. A pigment mark does not blanch, because melanin is not going anywhere. This is not a diagnosis and it does not replace an examination, but it is a genuinely informative thirty seconds, and it will often tell you which of the two waiting games you are in.

Atrophic means wasted or thinned. An atrophic acne scar is a depression: the repair process that closed the wound produced less collagen than was destroyed, so the surface sits below the surrounding skin. The great majority of true facial acne scarring is atrophic.

This is also the category where the reader gains the most from a bit of vocabulary, because atrophic scars are not one thing. The classification used across dermatology splits them into three shapes — ice-pick, boxcar and rolling — based on width, depth and the geometry of the walls. That split is not academic. Each shape has a different physical cause and therefore responds to a different physical intervention, and a great deal of the “I tried that and it did nothing” experience is a shape-treatment mismatch rather than a bad treatment.

Most faces have a mixture of all three, in different proportions, which is exactly why a single approach applied uniformly across a cheek tends to produce partial results. A clinician mapping a face will usually identify several types in the same region and plan accordingly.

An ice-pick scar is a narrow puncture — typically a couple of millimetres across at most — that runs steeply down into the dermis, tapering as it goes. In cross-section it is a V or a narrow column. On the skin it reads as a small dark dot, and it is frequently mistaken for a large pore or a blackhead, which is one reason people spend years attacking it with pore products that cannot possibly help.

The reason it is hard to treat is straightforward geometry. Anything that works by smoothing the surface — sanding, resurfacing, gentle collagen stimulation from above — would have to remove an enormous amount of surrounding healthy tissue to bring the general level down to the bottom of a deep narrow shaft. That is why treatments aimed at this shape tend to be targeted at the individual scar rather than the region: approaches that address the base of the shaft specifically, or that remove the shaft and close the gap. Those are procedures performed by clinicians, and the details belong in a consultation, not in an article.

The useful takeaway for a reader is simply this: if what you have is predominantly ice-pick, then broad surface treatments are a poor match for it, and knowing that in advance saves you from paying for the wrong category.

Boxcar scars are round or oval depressions with steep, clearly defined edges and a flat floor — like a shallow crater punched out of the skin. They are wider than ice-pick scars and can be shallow or deep. The name comes from the boxy cross-section: vertical walls rather than a taper.

Those sharp edges are the defining feature and the reason boxcars are so visible in certain light. An edge casts a hard shadow line. Anything that softens the edge or lifts the floor reduces the shadow, which is why the clinical approaches to boxcar scarring cluster around either resurfacing the rim so the transition becomes gradual, or physically removing or elevating the floor. Shallow boxcars generally respond better than deep ones, because there is less vertical distance to make up.

Boxcars are also the type most likely to look dramatically different between two photographs of the same face, because the shadow they throw depends entirely on the angle of the light. Hold that thought — it is the whole reason before-and-after galleries in this category deserve so little trust.

Rolling scars are the odd one out, because the problem is not really at the surface. They appear as broad, soft, undulating depressions with no sharp border — the skin looks slightly wavy rather than pitted, often across a wide area of cheek. The mechanism is fibrous tethering: bands of scar tissue anchor the underside of the dermis down to the deeper tissue below, pulling the surface into a dip.

This has an important implication. If the skin is being pulled down from beneath, then treatments that work on the surface are treating a symptom, and the tether is still there afterwards. The established clinical logic for rolling scars is to address the tether itself — which is a procedural intervention performed under the skin, not something available in a bottle. Combination approaches exist and are common, but the ordering matters, and that ordering is a clinical decision.

A practical clue: rolling scars change appearance markedly when you stretch the skin. Gently pulling the surrounding skin taut can flatten them substantially, because you are temporarily overcoming the tether. Sharply defined boxcars and ice-picks do not disappear under stretch in the same way. Again, this is not a diagnosis — but it is a real difference in behaviour that tells you the categories are physically distinct rather than just names on a chart.

Not all acne scarring is a hole. Some healing produces too much collagen instead of too little, leaving a raised, firm, often pink or darker lump. A hypertrophic scar stays within the original boundary of the injury. A keloid grows beyond it, sometimes considerably, and can continue expanding long after the acne itself has gone.

On the face these are less common than atrophic scars, but they are well recognised along the jawline, and they are much more frequent on the chest, shoulders and back. There is a strong individual predisposition — some people simply scar this way and know it from previous wounds, and there are recognised differences in susceptibility between skin types and family histories.

The reason to separate them out clearly is that the treatment direction is inverted. Atrophic scarring is treated by adding volume or stimulating collagen. Raised scarring is treated by reducing it. A protocol aimed at building collagen is, at best, pointless on a raised scar. This is also a category where self-treatment carries real risk: procedures that injure the skin can provoke more of exactly the response that caused the problem in someone predisposed to it. If what you have is raised rather than sunken, that fact alone is worth a professional opinion before anything else happens.

You cannot diagnose yourself and you should not try. But you can usually work out which broad bucket you are in, and that determines whether your next step is patience or an appointment. Three checks, in order.

1. Change the light. Flat frontal light — a window straight ahead, a ring light, a bathroom mirror lit from the front — erases texture almost completely, because texture is only visible through shadow. Move to a single light source off to one side, roughly level with your face, and look again in a mirror. Marks that vanish under side light and stay under flat light are colour. Marks that appear or deepen under side light are texture.
2. Press it. On a coloured mark, press gently and watch. If the colour momentarily disappears and floods back, it is vascular. If it does not change, it is pigment. Different problems, different timelines.
3. Feel it, then stretch it. Close your eyes and run a clean fingertip across the area. Flat means colour. A dent or a raised firmness means structural. If it is dented, stretch the skin gently: soft, wide dips that flatten out under tension behave like rolling scars, while sharp-edged pits that stay put behave like boxcar or ice-pick.

That is enough to have a much better conversation with a professional, and enough to stop you buying a fade cream for a crater. It is not enough to choose a treatment, and it is not a substitute for someone examining your skin in person under proper light.

Every appearance category has a photo problem. This one has the worst version of it, because both of the things being photographed are exactly the things a camera renders inconsistently.

Texture is shadow. Move the light from the side to the front and a field of boxcar scars becomes close to invisible without a single cell changing. Add a diffuser, or shoot on an overcast day instead of in a side-lit room, and the same thing happens. Nothing about this requires dishonesty; it is the default outcome of taking two photos in two different rooms.

Colour is white balance and exposure. A camera decides what counts as white in a scene and rebalances everything around that guess. Redness and brown pigment are precisely the tones that move most when that guess changes. Raise exposure slightly and low-contrast marks wash toward the surrounding skin tone. This is the same mechanism we walk through in detail for teeth in what actually changes in teeth whitening photos, and it applies unchanged to skin.

Then there is processing. Modern phones do not record a face, they reconstruct one, and skin smoothing is applied by default in many camera modes and by essentially every social platform. A pipeline whose explicit job is to remove skin texture is not a fair instrument for judging skin texture.

If you want a comparison you can believe, the discipline is the same as for any measurement: one room, one light source with the others switched off, the light in the same position, the same distance and camera height, locked white balance and exposure, no filters and no beauty mode, and the same expression. Add a plain white card in frame as a reference. Then take the tracking shot deliberately side-lit, because you want to see texture, not flatter it.

We should be direct about this, because the topic invites exactly the misunderstanding we want to avoid. RealSmile derives 17 facial-geometry metrics from 68 facial landmarks and turns them into a validated Face Score percentile. On /analyze it also reads smile genuineness — whether the orbicularis oculi is engaged, which is the Duchenne marker that separates a felt smile from a posed one.

Among those metrics is Skin Clarity. It is a read on how even the skin appears in the photograph you uploaded. That is the whole claim. It cannot tell pigment from a dent, it cannot tell a healing spot from a scar, it cannot see below the surface, and it has no diagnostic capability of any kind. It will not tell you whether you have boxcar scarring. It cannot distinguish post-inflammatory hyperpigmentation from an atrophic scar, because in a flat-lit photograph those two things can produce a nearly identical patch of pixels. The same applies to the other appearance-adjacent metrics: Brow Arch and Brow-Eye Proximity are geometry, not health.

There is a second limit worth stating even more plainly, and it comes from our own data rather than from the literature. When we looked at people rescanning themselves within a single session, the score for one face routinely moved by several points across shots. The face did not change. The lighting, camera height, head angle and expression did, and that was enough to shift landmark placement and everything computed from it. If a geometric measurement is that sensitive to capture conditions, a skin-appearance read — which depends on light direction, white balance and processing, none of which affect geometry — is more sensitive still.

So the honest framing of a Skin Clarity number is: this is a description of a photograph, taken under conditions you controlled or failed to control. It is useful as a repeated measure if you keep those conditions fixed. It is not evidence about your skin, and it is not a screening tool. The 17-metric scan and the face rating tool both push you toward even, directional light and a level camera for that reason.

We are not qualified to tell anyone what to do, and we will not. What we can usefully do is describe the shape of the consultation, so that it is less opaque when you have one and so you can tell whether you are getting a real assessment.

A proper assessment generally covers: whether the acne is still active, because scar treatment is usually sequenced after the underlying condition is controlled; which categories of mark are actually present and in what proportions, examined under directional light and often with the skin stretched; the depth and wall geometry of any atrophic scars; whether anything is raised rather than depressed; your skin type and pigment-response history, because that changes the risk profile of several interventions; whether you scar or keloid readily; your history of previous treatments and how the skin reacted; and any medication history relevant to how the skin heals, which is a specific and important question for anyone who has taken systemic acne treatment.

Only after that does a plan get built, and a realistic plan is usually a combination rather than one procedure, staged over a considerable period, with an expected outcome described as improvement rather than removal. Anyone offering a single treatment that erases all acne scarring is describing an outcome the field does not claim.

Retinoids come up constantly in this conversation, and it is worth explaining what they are without turning that into a recommendation. Retinoids are vitamin A derivatives that act on receptors in skin cells, influencing how quickly cells in the epidermis turn over, how pigment is handled, and how the dermis produces collagen. That combination is why they appear in discussions of both pigmentation and texture.

Here is the part that matters practically. Tretinoin is a prescription drug in most jurisdictions. Other retinoid molecules sit at different regulatory tiers in different countries — some over the counter, some not. They are not interchangeable, they differ in strength and in tolerability, they commonly cause irritation particularly at the start, and irritation is itself an inflammatory event, which in a person prone to post-inflammatory pigmentation is the specific thing you are trying to avoid. There are also situations, including pregnancy, where they are contraindicated.

So the mechanism is legitimately interesting and the class is genuinely central to dermatological practice, and neither of those facts makes it appropriate for a face-measurement website to tell you to start one, at what strength, or how often. That is a conversation with a doctor or pharmacist who can see your skin, knows your history, and is accountable for the answer. The same reasoning applies to every other active ingredient people ask about here, and it is why the skincare routine guide stays on habits and measurement rather than prescribing molecules.

This is unglamorous and it is also true. Every intervention discussed above is an attempt to partially reverse something. The largest available effect is not creating it.

Two things are consistently stressed. The first is treating active inflammatory acne early and properly rather than waiting it out, because scarring risk relates to the depth and duration of inflammation, and deep nodular lesions left to run their course are the ones most associated with permanent textural change. The second is not manipulating lesions. Squeezing and picking extend the inflammatory event, drive material deeper into the dermis, and are a well-recognised route from a spot that would have healed cleanly to a mark that will not.

Neither of those is advice about your specific case. Both are simply the direction the entire field points, and they are cheaper than every procedure on the list.

Colour marks fade on the scale of months, sometimes many months, and deeper pigment slower than superficial pigment. Because that fading happens anyway, it is the perfect environment for a product to take credit for the passage of time. If you started something four months ago and things look better, you have two hypotheses and no way to separate them without a controlled baseline.

Textural scars do not follow that pattern. They are stable. If a dent has been there a year, it will be there next year. The upside of that stability is that it makes evaluation clean: any change in a textural scar is attributable to whatever you did, provided your photographs are honest. The downside is that it removes the comfortable option of waiting.

Procedural treatments also have their own delays. Collagen remodelling is slow, so approaches that work by stimulating the dermis are typically judged over months rather than weeks, and are often described in terms of a course rather than a session. That is worth knowing before you evaluate anything after two weeks and conclude it failed.

The practical protection against all of this is a baseline. One set of photographs, side-lit, fixed conditions, taken before you start anything, kept somewhere you will find them. Memory for skin appearance is poor in both directions — people fail to notice real improvement and also convince themselves of improvement that is not there. A fixed reference beats both.

These tend to produce specific answers, and the quality of the answer tells you something about the quality of the assessment.

• Which of what I have is pigment or redness, and which is actual textural scarring? How are you telling them apart?
• Of the textural scarring, what types are present and roughly in what proportion?
• Is my acne controlled enough to be treating scars yet, or should that be sequenced first?
• Given my skin type, what is the risk that this treatment causes pigment change of its own, and how is that managed?
• What proportion of improvement is realistic here, over how many sessions and how long?
• What is the downtime, what does the healing look like, and what would count as a complication I should call about?
• How will the baseline be recorded, and will the follow-up photographs use the same lighting and camera position?

That last question is not a throwaway. If nobody records a controlled baseline, the only comparison available later is between a new photograph and your recollection, and that comparison is worthless. A practice that photographs consistently is telling you something about how it evaluates its own work.

Flat and coloured is not a scar. It is pigment or it is vessels, it usually improves with time and photoprotection, and the main job is not adding more of it while you wait. Dented or raised is a scar. It will not fade, the shape determines which intervention makes physical sense, and the assessment has to be done in person by someone who can look at your skin under proper light and stretch it with their fingers.

Between those two facts sits almost everything worth knowing. Add one habit — a side-lit baseline photograph under conditions you can reproduce — and you have the ability to tell whether anything you do afterwards actually worked, which is more than most people in this category ever have.