Welcome back to the journal review. This is the June 2026 issue of Dermatologic Surgery, and we've got four articles on the docket this time — a retrospective look at facelifts after filler, a pediatric nail algorithm, a meta-analysis pitting superficial radiation against Mohs, and a machine-learning scheduling tool. Let's get into it.

First up is an original article asking a question that comes up constantly in consult rooms: does prior treatment with facial injectables increase the risk of complications from rhytidectomy — that is, facelift surgery. This is a retrospective chart review out of a single plastic surgery practice in Toronto, looking at every deep-plane facelift performed over a fifteen-month stretch. The clinical gap here is real — there's been a documented rise in both facelift volume and filler volume over the past several years, so more and more patients are showing up for surgical consultation with a filler history, and yet there was essentially no evidence-based data on whether that history matters surgically. Anecdotally, some surgeons have reported that revision facelifts are harder to dissect in patients with heavy prior filler use, and in one survey the authors cite, nearly forty percent of plastic surgeons believed panfacial filler history increased surgical risk. So this was purely an assumption looking for data.

Methodologically, this is about as straightforward a retrospective design as you'll see — one surgeon, one technique, one consecutive case series, split into two groups: those with a history of hyaluronic acid, calcium hydroxylapatite, and poly-L-lactic acid injectables, and those without. The authors don't spend much time justifying the retrospective design, but it's the obvious first step for a question like this — you can't ethically or practically randomize people to filler versus no filler before elective facelift surgery, so a chart review of real-world practice is the sensible starting point, even though, as we'll get to, it comes with real limitations around recall.

On to results. This was a modest cohort, one hundred six patients, mean age around sixty-one, overwhelmingly female. Just over half — fifty-seven percent — had a history of injectables before surgery. Overall, about one in four patients had some kind of complication, and when you split the groups, sixteen percent of the prior-injectable group had a complication compared to eight percent of the no-prior-injectable group. That's a doubling in raw rate, which sounds like it should matter, but statistically it was not significant, and the authors are explicit about that. Surgical time was essentially identical between groups too — a little under five hours either way, also not a significant difference. All the complications recorded — things like transient tissue ischemia, small hematomas, localized infections, dehiscence, and granulomas — were characterized as mild and fully resolved, none requiring a return to the operating room.

The discussion is appropriately measured. The authors point out that all these cases were done using a deep-plane technique, which dissects underneath the SMAS in a relatively avascular plane with minimal skin undermining — so even if filler had created some fibrosis or tissue distortion more superficially, it may simply not be in the surgical path for this particular technique. They also raise the possibility of a dose effect — maybe these patients just hadn't received enough cumulative filler volume to matter, and higher-volume panfacial filler patients might behave differently. The honest limitations are worth flagging: this is a single surgeon, single technique, modest sample size, and critically, it relies on patient-reported injectable history — where exactly they were injected, with what, and when is often incomplete or unreliable, so misclassification is a real possibility. The subgroup analysis by complication type was also underpowered given how few events there were in each category.

Practically, here's the takeaway — this is reassuring, useful counseling data, but it is not yet practice-changing in the sense of proving safety definitively. What you can tell a patient is that the best available evidence so far shows no statistically significant increase in complications after deep-plane facelift in patients with a filler history, and no increase in operative time or difficulty. What you cannot yet say is that this generalizes to different techniques, higher filler volumes, non-HA permanent fillers, or multiple surgeons. Consider this hypothesis-generating first evidence rather than a closed question.

Next is a review article — a practical treatment algorithm for pediatric ingrown toenails, put together by a multinational group of nail specialists. This isn't a study with its own new data; it's a synthesis of existing literature translated into a decision tree, so I'll walk through it the way a review deserves — what the evidence shows and how they've organized it clinically, rather than forcing methods and limitations onto it.

The clinical problem is one most of us see referred in from pediatrics or podiatry more than we generate ourselves: ingrown toenails in children and adolescents, where anesthesia tolerance, compliance, and the risk of permanently narrowing a growing nail plate all complicate decision-making in ways that don't apply the same way in adults. The authors use the Heifetz staging system — stage one is just inflammation and erythema of the lateral fold, stage two adds granulation tissue and drainage, stage three is chronic hypertrophy of the fold from long-standing inflammation — and they build their algorithm around that staging, layered with age.

On the evidence itself: conservative measures — taping, cotton pledgets, nail bracing — are painless and well tolerated, but recurrence is all over the map, roughly five percent up to about one in four, depending on technique, with wound closure strips performing best and cotton pledgets performing worst. One large study of taping techniques in over a thousand pediatric cases reported success rates in the mid-nineties percent range across all three stages, but treatment courses can run for months, which is a real problem when a child is in pain or has poor tolerance for prolonged conservative care. On the surgical side, nail edge excision alone had recurrence as high as a third of cases, but combining it with chemical matricectomy dropped that down to around twelve percent — clearly the more effective surgical combination. Techniques like the Vandenbos procedure, which excises the hypertrophic lateral folds rather than narrowing the nail plate, reported no recurrences in the cited literature, while complete nail avulsion alone was actually the worst performer among surgical options, with recurrence over half.

The algorithm itself is really the practical payload here, and it's age-stratified in a clinically meaningful way. Under age ten, the guidance is to avoid chemical matricectomy and plate-narrowing procedures altogether, because of the risk of permanent nail distortion in a still-developing plate — instead favor debulking the hypertrophic soft tissue with a sharp curette, using techniques like Vandenbos, Super-U, or Dubois depending on which folds are involved. Over age ten, chemical matricectomy with phenol, trichloroacetic acid, or sodium hydroxide becomes reasonable for stage two and three disease, paired with curettage of granulation tissue or hypertrophic tissue removal as needed. Anesthesia and setting recommendations follow the same age split — younger children generally need a hospital setting with sedation or general anesthesia coordinated with anesthesiology, while older kids can typically be managed with a local digital block in clinic.

For practical purposes, this is a genuinely useful reference algorithm rather than a practice-changing trial — think of it as a curated flowchart you'd want on hand before your next pediatric nail referral, particularly the age-ten cutoff for avoiding matricectomy and the reminder to actually assess gait and foot structure as a contributor to recurrence, which the authors emphasize is frequently skipped.

Third is a systematic review and meta-analysis comparing local recurrence rates between superficial radiation therapy and Mohs micrographic surgery for nonmelanoma skin cancer — squarely in our wheelhouse, since this bears directly on how we counsel patients who come in asking about the noninvasive alternative they saw advertised.

The background problem is straightforward: SRT use has grown alongside more accessible in-office kilovoltage machines, and some single-institution SRT series have reported five-year local recurrence in the four to six percent range, fueling patient interest. But there had never been a systematic, pooled comparison against Mohs recurrence data. The authors ran a PRISMA-compliant systematic review across four major databases through November of 2024, ultimately pooling twenty-six studies — nine reporting on SRT, seventeen on Mohs — using a random-effects meta-analysis model. That random-effects choice matters methodologically: given how clinically heterogeneous these studies were — different tumor sites, sizes, stages, follow-up durations — a random-effects model is the appropriate, more conservative approach rather than assuming one true fixed effect across all studies. They also ran a formal risk-of-bias assessment using the Cochrane ROBINS-I tool and graded evidence certainty using GRADE, and did a sensitivity analysis dropping each study one at a time to confirm no single study was driving the pooled result — which it wasn't.

The results are about as clean a signal as you'll get in a meta-analysis of this kind. Pooling almost eight thousand SRT cases and just over ten thousand Mohs cases, the weighted local recurrence rate for Mohs came out at under two percent, compared to roughly six percent for SRT — a difference of about twelve percentage points, and this was statistically significant. Follow-up times were comparable between groups, both averaging a bit over four years, so this isn't an artifact of Mohs simply being followed for less time.

The authors' discussion reinforces what you'd expect: Mohs remains the gold standard for tumors meeting Appropriate Use Criteria, in large part because it provides complete circumferential and deep margin assessment with histologic confirmation of clearance, something SRT fundamentally cannot offer since there's no tissue processing involved. They do carve out a reasonable niche for SRT — elderly or frail patients, those with a high likelihood of ending up with a chronic nonhealing wound after surgery, medical contraindications to surgery, or patients who simply refuse an incisional procedure. But they're clear that outside those scenarios, the recurrence data favor Mohs. The honest limitation, which the authors themselves flag prominently, is substantial heterogeneity — statistically very high across both arms — driven by uncontrolled differences in tumor histology, size, location, stage, and surgical or radiation technique across studies, plus the observation that SRT studies tended to include more superficial, lower-risk tumors than the Mohs studies, which could bias the comparison in SRT's favor rather than against it.

Practically, this is a useful, citable number for patient counseling — you can now say that pooled data across more than twenty studies shows roughly a three-fold higher local recurrence rate with SRT compared to Mohs, without reaching for hedgy language. It's not new practice guidance — nobody is changing AUC-based indications because of this — but it is a solid piece of evidence to put in front of a patient who's leaning toward SRT for a tumor that clearly meets Mohs criteria.

Last is the CLASS project — an original proof-of-concept study out of Mayo Clinic applying machine learning to Mohs surgical scheduling, which is a refreshingly operational question rather than a clinical outcomes one.

The background problem is one every busy Mohs practice recognizes: case complexity is unpredictable at the time of scheduling, and getting it wrong — putting a complex midface tumor at the end of the day, for instance — cascades into overtime, staff fatigue, and long patient waits. Existing complexity grading systems tend to be practice-specific and don't generalize well. So the authors asked whether machine learning, trained on data already available at the time of referral, could predict two surrogates of complexity: number of Mohs stages needed, and whether reconstruction would end up being complex, meaning a flap, graft, or referred closure.

Methodologically, this drew on a large multi-site dataset — over twenty thousand Mohs cases across six practices and fourteen surgeons within the Mayo system from 2018 to 2023, split into a training and validation set of about eighteen thousand cases and a held-out test set of just over two thousand. They used an automated machine learning framework, testing multiple algorithm types, with ten-fold cross-validation, and restricted the input features to only four variables — sex, age, histopathologic diagnosis, and anatomic location — because those are the only data points reliably available at the time a case is actually being scheduled, before the surgery itself happens. That's a smart, deliberate constraint: the model had to be useful prospectively, not just retrospectively descriptive. They paired this with a separate qualitative workflow analysis conducted by health systems engineers observing the practice in person, and then combined both data streams into fixed scheduling rules — a nice example of triangulating a statistical model with human operational expertise rather than trusting the algorithm in isolation.

On results, the model predicting complex reconstruction performed quite well, with an area under the curve around zero-point-eight-three — that's a clinically meaningful, reasonably strong discriminative model. The model predicting number of stages performed much more modestly, with an area under the curve around zero-point-six-two to zero-point-six-three, which is only slightly better than chance and reflects real difficulty predicting stage count from limited preoperative variables alone — not surprising, since stage count is influenced by things like actual tumor extent and subclinical spread that simply aren't knowable before the case starts. Across both models, anatomic location was by far the most important predictive feature, with midface sites — nose, lip, eyelid, anogenital region — consistently pushing predictions toward more stages and more complex reconstruction, while trunk, neck, and extremity sites pushed the other way. A couple of specific diagnoses, dermatofibrosarcoma protuberans and extramammary Paget's disease, also nudged predictions toward more stages, which lines up with what any of us would expect clinically given their tendency toward subclinical extension.

The authors are appropriately modest in their framing — this is explicitly a proof-of-concept, not a finished, deployable product. The stage-prediction model in particular isn't accurate enough on its own to drive scheduling decisions, and the whole framework was built and validated within one health system's data, so generalizability to other practices with different referral patterns, staffing models, or patient populations is untested. There's also the inherent messiness they acknowledge in matching Mohs operative notes to pathology reports using a fuzzy algorithm rather than a direct link in the electronic record, which introduces some noise into the whole pipeline.

Practically, this one is interesting rather than practice-changing for now — it's not a tool you can go implement tomorrow. But conceptually, it validates something most experienced Mohs surgeons already intuit: anatomic location, especially midface, is the dominant driver of case complexity and time. The real signal here is directional — it suggests that even simple, readily available scheduling-time variables can meaningfully flag high-complexity cases for better slotting, and it's a reasonable bet that more refined versions of this kind of tool, especially ones that improve stage prediction, will start showing up as scheduling aids in larger practices over the next few years.

That wraps up this June 2026 issue — a reassuring signal on facelift safety after filler, a clear age-stratified roadmap for pediatric ingrown toenails, solid pooled evidence reaffirming Mohs' recurrence advantage over superficial radiation, and an early but promising look at machine-learning-assisted scheduling. Thanks for listening, and we'll see you next month.