Welcome back to this month's journal walkthrough — the Journal of the American Academy of Dermatology, January twenty twenty-six issue. We've got four pieces to get through this time: an original cohort study questioning whether adding chemotherapy to adjuvant radiation actually helps in high-risk cutaneous squamous cell carcinoma, an editor's letter on topical timolol and wound healing, a response letter critiquing a biologics-and-skin-cancer cohort study, and a National Cancer Data Base analysis on what predicts metastatic sebaceous carcinoma. Let's get into it.

First up is a dual-center retrospective cohort study asking a question a lot of us face in tumor board: once you've decided a cutaneous squamous cell carcinoma needs adjuvant radiation therapy, does bolting on a chemosensitizer — cisplatin, carboplatin, or cetuximab — actually buy your patient anything? The background here is thin evidence. There's exactly one prospective trial, the TROG oh-five-oh-one study, looking at platinum-based chemoradiation versus radiation alone in head and neck disease, and it found no benefit in overall survival, disease-free survival, or locoregional recurrence. But the authors point out real weaknesses in that trial — it lumped in some lower-risk tumors as advanced disease, and it used carboplatin alone, which is generally considered the weaker platinum agent compared to cisplatin. Meanwhile, smaller cetuximab series look more favorable, but they tend to include lower-risk tumors or have no comparison arm at all. So this is a genuine gap: nobody has cleanly compared adjuvant chemoradiation against radiation alone in a real-world, higher-risk cohort with a proper comparator.

Methodologically, this was two centers, all squamous cell carcinomas that received adjuvant radiation over a multi-year window with at least three years of follow-up, looking at three endpoints — local recurrence, locoregional recurrence, and disease-free survival. Here's the part worth pausing on for methodology's sake: because this is retrospective, the patients who got chemosensitizer added were not going to look like the patients who didn't — clinicians reach for chemo in the scarier tumors. So the authors used inverse probability of treatment weighting to try to balance the two groups on things like nodal disease, margin status, size, and depth before running the actual outcome models. And for the recurrence endpoints, they used Fine-Gray competing risk modeling rather than a standard Cox model — which is the right call, because death competes with recurrence as an outcome, and ignoring that competing risk can inflate apparent recurrence rates. Disease-free survival, which doesn't have that same competing-risk problem, was modeled with a standard Cox proportional hazards approach.

Now, even after weighting, the chemoradiation group still looked meaningfully higher risk — more nodal disease at presentation, more recurrent tumors, more positive margins, more advanced stage. That imbalance matters for how you read the results. About two hundred ninety-four patients total were analyzed, the vast majority getting radiation alone and a smaller group of around thirty getting chemoradiation, split roughly with about half getting cisplatin, roughly a third getting cetuximab, and a handful getting carboplatin or even a second agent like paclitaxel layered on. Bottom line: adding a chemosensitizer did not improve local recurrence, did not improve locoregional recurrence, and did not improve disease-free survival. None of the differences reached statistical significance, and the actual recurrence numbers were low and similar between groups — roughly six percent versus two percent for local recurrence at five years, and comparable locoregional recurrence whether or not nodal disease was present at baseline. Three-year disease-free survival came in at about sixty-two percent with radiation alone versus sixty percent with chemoradiation — essentially a coin flip, not a clinically meaningful gap.

The honest discussion here is nuanced. The authors don't conclude that chemosensitization is useless — they note that because the chemoradiation group was so much higher risk even after statistical balancing, it's entirely possible those patients would have done worse without the chemotherapy, and the weighting simply couldn't fully erase that residual confounding by indication. That's a real limitation of propensity methods in general — they balance what you measure, not what you don't. Add to that a small chemoradiation cohort of around thirty patients, and the study is underpowered to detect anything but a large effect. So where does this leave you practically? This is not practice-changing in the sense of telling you to stop using chemosensitizers — it's better read as confirmatory null data reinforcing what the one prior prospective trial suggested, that routine addition of platinum agents or cetuximab to adjuvant radiation lacks a demonstrated survival or recurrence benefit in this population. The decision remains individualized, and with adjuvant immunotherapy now emerging as a real competitor in this space, chemotherapy is increasingly going to be reserved for patients who can't tolerate or don't qualify for immunotherapy — which is exactly the framing the authors land on.

Next is a letter from the editor, Dirk Elston, responding to a randomized trial elsewhere in this same issue looking at topical timolol gel for wound healing. This is a commentary, not original data, so think of it as expert synthesis rather than a new methods-and-results paper. The trial in question, by Anabaraonye and colleagues, randomized patients with open surgical wounds to zero-point-two-five percent timolol gel versus white petrolatum, and found timolol did not accelerate re-epithelialization. Elston flags an important imbalance in that trial — the timolol group was younger and had less diabetes, which should favor timolol, but they also had more peripheral vascular disease and were more often current smokers, which cuts the other way — so the net effect on interpretation is genuinely ambiguous. He also makes a conceptual point worth remembering: white petrolatum itself is not an inert control, it has its own documented positive effect on wound healing, and in his own practice he's never used timolol alone without pairing it with a base like petrolatum. The bigger point, though, is a distinction in wound biology — this trial studied active, healing surgical wounds, which behave very differently from chronic, static, nonhealing ulcers, and he cautions against generalizing a negative finding in one context to the other.

Where he thinks timolol does have a role, based on the broader literature, is in reducing excess granulation tissue after surgery — similar to its known effect on the proliferative phase of infantile hemangiomas — and in reducing vascularity and erythema, as shown in a small trial of post-mammoplasty scars. For chronic wounds specifically, he cites a meta-analysis of four studies in refractory venous ulcers showing a significant reduction in ulcer area at two weeks, though he's careful to note the caveats the original authors raised — small samples, no blinding, only four studies feeding the pooled estimate. A separate review of two randomized trials and ten observational studies similarly leaned favorable but carried the same blinding and sample-size limitations. He also references a study using timolol maleate for wounds of varying etiology that had failed prior treatment, describing improved healing, and a post hoc analysis from a VA-based diabetic foot ulcer trial showing timolol reduced transepidermal water loss in healed ulcers, though re-ulceration rates were unchanged — and he's explicit that post hoc analyses are prone to bias. His own bottom line, stated plainly: in his practice, topical timolol — usually paired with gentamicin ointment or petrolatum — is useful for reducing excess granulation tissue and promoting re-epithelialization in chronic, static, nonhealing wounds. He does not currently use it for actively healing postsurgical wounds, which aligns with the negative trial this letter is responding to. So take this as reinforcing a fairly narrow, specific niche for topical timolol — chronic wound management, not routine postsurgical wound care.

Third is a response letter — essentially a methodologic critique — addressing a prior retrospective cohort study by Lauck and colleagues that looked at cutaneous malignancy risk after biologic therapy for inflammatory disease, using an active-comparator design with propensity score matching. The letter writers commend the study's overall contribution but raise two specific concerns worth flagging for your own critical reading of pharmacoepidemiology papers like this. First, the original study didn't account for combination therapy — in rheumatology, TNF inhibitors are frequently combined with methotrexate, and prior meta-analyses and large observational studies have suggested that combination may further raise non-melanoma skin cancer risk beyond either agent alone, a possible synergistic effect. The letter argues that combination regimens should ideally be modeled either as their own separate exposure category or as an explicit covariate, otherwise you risk over- or under-estimating the risk attributable to the biologic itself. Second, and probably the more fundamental issue, is that the database lacked any measure of ultraviolet exposure — far and away the dominant environmental risk factor for skin cancer, and one that varies enormously by occupation, geography, and sun habits. Layer onto that a study population that was predominantly White and drawn from Nordic or North American health systems with no stratified analysis by race or region, and the letter's point is that absolute risk estimates from this kind of cohort shouldn't be casually extrapolated to more diverse patient populations, where baseline skin cancer risk and UV sensitivity differ substantially. This is a letter, so there's no new data here — just a call for future studies to model combination therapy explicitly and to stratify by race and geography, which is a fair methodological ask any time you're leaning on a cohort like this to counsel patients on biologic selection.

Last is a retrospective cohort study using the National Cancer Data Base to identify risk factors for regional or distant metastatic sebaceous carcinoma at presentation. The clinical backdrop: sentinel lymph node biopsy is already recommended for periocular sebaceous carcinoma once it reaches T2c or beyond — full-thickness eyelid tumors over ten but under twenty millimeters. But for extraocular disease, current guidelines actually recommend against metastatic workup, based on older Surveillance, Epidemiology, and End Results data suggesting only about a one percent metastasis rate in extraocular head and neck tumors — data that, notably, never compared that group against trunk and extremity disease. This study set out to fill that gap using a much larger, more complete national dataset.

Methodologically, they pulled every case with the appropriate histology code that had complete data on primary site, stage at diagnosis, surgery type, and tumor size — worth noting they emphasize completeness here, because a prior comparable SEER-based study was missing tumor size data on nearly two-thirds of its cases, so this cohort is genuinely a step up in data quality even though the National Cancer Data Base carries its own selection bias toward larger, accredited cancer centers. They ran univariate chi-square screening followed by multivariate logistic regression to isolate independent predictors.

Out of about thirteen hundred fifty cases, the large majority — around eighty-five percent — were extraocular, and strikingly, when you looked at just the metastatic cases, over three-quarters of those were also extraocular. On multivariate analysis, four factors independently predicted metastatic disease at presentation: tumor size two centimeters or larger compared to under one centimeter carried roughly an eight-fold increase in odds, and even the one-to-two centimeter range carried about a two-fold increase, both statistically significant. High histologic grade compared to low grade carried about a five-fold increase in odds. And critically, anatomic location mattered independent of size and grade — periocular tumors carried roughly a three-fold increased odds of metastasis compared to trunk and extremity disease, and non-periocular head and neck tumors carried about a two-and-a-half-fold increase, also compared to trunk and extremities.

The discussion here directly challenges current practice guidance. Since most of the metastatic cases in this cohort were extraocular head and neck tumors — a group current guidelines explicitly say doesn't need metastatic workup — the authors argue that recommendation may need revisiting, particularly for larger or higher-grade extraocular head and neck lesions. The honest limitations are real: this is a Commission on Cancer-accredited hospital dataset, so there's inherent selection bias toward more complex referral cases; histologic grade was unknown in a substantial chunk of both metastatic and localized cases; and the absolute number of metastatic cases, while better annotated than prior series, is still only in the eighties out of the full cohort, which limits precision on some of these subgroup estimates.

Practically, here's the distinction to draw. This is not a guideline change on the page — sentinel lymph node biopsy criteria for periocular disease remain what they are — but it's a meaningful evidence-based nudge that extraocular head and neck sebaceous carcinoma, particularly when it's two centimeters or larger or high grade, probably deserves more metastatic workup consideration than the current "extraocular equals low risk, skip the workup" framework suggests. I'd call that clinically actionable food for thought the next time you're staging one of these tumors, even if it hasn't yet been codified into formal guidelines.

That wraps this month's four articles — a null result on chemosensitized radiation for high-risk squamous cell carcinoma, a nuanced editorial take on where topical timolol actually helps versus doesn't, a methodologic critique worth keeping in mind when you read biologic-and-cancer-risk cohort studies, and new data pushing back on the assumption that extraocular sebaceous carcinoma is uniformly low risk. Thanks for listening, and I'll see you next month.