Why this matters
Cutaneous angiosarcoma presents a difficult combination of local recurrence risk and early distant dissemination. This prospective phase II study evaluated a treatment sequence intended to address both: induction paclitaxel, followed by concurrent paclitaxel and radiotherapy, then surgery when feasible.
The reported disease control is encouraging, but the treatment burden is substantial. More than half of evaluable patients experienced severe acute toxicity, and nearly one-third experienced severe late toxicity.
The clinical story is therefore not simply a high response rate. It is whether an intensive multimodality strategy can deliver sufficiently durable control to justify its toxicity.
Study design
This prospective, single-arm phase II trial enrolled adults with localized, histologically confirmed cutaneous angiosarcoma and ECOG performance status 0-1.
Treatment began with six weeks of induction paclitaxel at 80 mg/m² weekly. Patients then received concurrent weekly paclitaxel at the same dose with external beam radiotherapy, using 50 Gy in 25 fractions or 50.4 Gy in 28 fractions.
Surgery followed chemoradiotherapy when feasible. An additional radiation boost and hyperthermia were permitted at physician discretion.
The primary endpoint was two-year progression-free survival. Secondary endpoints included overall survival, pathological complete response, and grade 3 or higher acute and late treatment-related toxicity.
Twenty patients were enrolled, with 19 evaluable. Median age was 71 years. Eight patients had head and neck disease, and 11 had breast or chest wall disease. Median follow-up was 25 months.
Key results
The primary endpoint estimate was encouraging.
Estimated two-year PFS was 82.5%. The presentation separately reported:
| Two-year outcome | Estimate |
|---|---|
| Progression-free survival | 82.5% |
| Local control | 94.4% |
| Locoregional control | 82.5% |
| Overall survival | 87.4% |
Local control and locoregional control should not be used interchangeably. The presentation reported only one in-field recurrence among the head and neck cases.
Most patients proceeded to surgery.
Surgical resection was performed in 18 of 19 evaluable patients. Negative margins were reported in 17 of the 18 resections.
Pathological complete response occurred in 10 patients, representing 56% of those undergoing surgery or 53% of the evaluable cohort. These are two descriptions of the same response count, not different efficacy results.
Treatment completion required frequent modifications.
All 19 evaluable patients completed concurrent chemoradiotherapy, but two required RT dose reductions. The presentation reported chemotherapy dose reductions in 10 patients (53%), while the abstract reported chemotherapy breaks in 11 patients (58%).
Two patients received an RT boost, and four received hyperthermia. The results therefore reflect a treatment strategy with some individualized components rather than an entirely uniform regimen.
Severe toxicity was frequent.
Grade 3 or higher acute treatment-related toxicity occurred in 11 patients (58%), and grade 3 or higher late toxicity in six patients (32%). Radiation dermatitis was the most common toxicity.
The presentation added clinically important detail: two patients experienced grade 4 leukopenia, and one developed grade 4 skin ulceration approximately two years after treatment.
There were no treatment-related deaths. Nevertheless, the absence of fatal toxicity should not be interpreted as evidence of a low-toxicity regimen.
Interpretation
The strongest contribution is prospective evidence that this paclitaxel-based treatment sequence can produce substantial pathological response and encouraging disease control in localized cutaneous angiosarcoma.
However, the outcomes belong to the whole multimodality strategy. Almost every evaluable patient underwent surgery. These results do not establish chemoradiotherapy as an alternative to resection, and they cannot isolate the contribution of induction paclitaxel from concurrent treatment or surgery.
The pathological response finding is relevant but should remain in proportion. Ten complete responses in resection specimens demonstrate treatment activity. They do not establish that surgery could have been omitted or that pCR will translate into a durable survival advantage.
The site-specific results also suggest that the pooled estimate conceals clinically relevant heterogeneity. The presentation reported more favorable outcomes in breast/chest wall disease than in head and neck disease. With only 11 and eight patients, respectively, these are descriptive observations rather than reliable evidence for selecting treatment by anatomical site.
Toxicity is central to judging the strategy. High completion rates coexist with frequent chemotherapy modifications, severe dermatitis, and a late grade 4 ulceration. Completion alone therefore does not establish that treatment was straightforward or well tolerated without qualification.
The investigators compared their outcomes favorably with historical cohorts. That provides context, but not proof of superiority. Differences in patient selection, disease site, surgery, and follow-up can influence cross-study comparisons.
Limitations
Only 19 patients were evaluable, and there was no concurrent control group. The study cannot establish improved survival or disease control over another treatment strategy.
The reported material does not provide confidence intervals around the main survival estimates or the prespecified statistical benchmark for declaring the primary endpoint successful. The two-year PFS estimate should therefore not be presented as a demonstrated comparative benefit.
Median follow-up was 25 months, leaving uncertainty about longer-term recurrence and toxicity. The late ulceration reported around two years reinforces the importance of continued observation.
The cohort combined head and neck with breast/chest wall disease. Optional boosts, hyperthermia, and frequent dose modifications further limit attribution of outcomes to one standardized regimen.