Three of this week's four trials added something. SOLARIS added vitamin D to first-line chemotherapy in unselected colorectal cancer and got 11.8 months against 10.3. SKYSCRAPER-03 added an anti-TIGIT antibody to consolidation in unselected stage III lung cancer and got a hazard ratio of 0.96. The osimertinib trial added the same chemotherapy FLAURA2 had already tested, but only in patients carrying a TP53 co-mutation, and more than doubled progression-free survival.
The difference was not the drug. It was the denominator. TP53 status sits on most sequencing reports we already order, and most of us glance past it. Vitamin D may yet work in left-sided tumours; that subgroup has now appeared three times. Selection is the cheapest intervention available, and it is the one we keep skipping.
Every practice-changing trial of the week, distilled to a single call. Act on it, or wait.
Every verdict weighs four things: the trial phase, whether the endpoint was mature, the size of the effect, and what it was measured against. Signal is reserved for randomised phase 3 evidence strong enough to act on now. Watch means the result is real but not yet enough to change practice, and most readouts land here.
PSMAddition. [177Lu]Lu-PSMA-617 at 7.4 GBq every 6 weeks for up to six cycles, added to androgen deprivation therapy plus an androgen receptor pathway inhibitor, versus ADT plus ARPI alone, in PSMA-positive metastatic androgen pathway modulator-naive or sensitive prostate cancer, randomised phase 3, 169 sites in 20 countries, n=1144, second interim analysis at 311 events. Radiographic progression or death occurred in 139 of 572 (24 percent) versus 172 of 572 (30 percent), hazard ratio 0.72 (95% CI 0.58 to 0.90, p=0.0021), meeting the primary endpoint. Time to PSA progression favoured the radioligand strongly (hazard ratio 0.42). Median radiographic progression-free survival was not reached in either arm, and overall survival did not differ (hazard ratio 0.84, 95% CI 0.63 to 1.13, p=0.13) at 184 deaths. Dry mouth affected 46 versus 4 percent, all grade 1 or 2. Composite time-to-worsening hazard ratios for FACT-P, EQ-5D-5L and pain all sat above 1.0, though every confidence interval crossed it. The FDA approved this regimen on 31 July on the strength of these data, so Watch is a deliberate disagreement with the regulator rather than a wait for one. A genuine positive phase 3 that has so far bought delayed imaging progression, not longer or better life. In a disease measured in years, six cycles of radioligand on top of standard doublet therapy needs a survival curve before it becomes routine, and the regulatory filing cites a later cutoff with overall survival still crossing 1 (hazard ratio 0.80, 95% CI 0.63 to 1.01).
SOLARIS (Alliance A021703). High-dose vitamin D3 (8000 IU daily for 14 days, then 4000 IU daily) versus standard-dose (400 IU daily), added to mFOLFOX6 or FOLFIRI plus bevacizumab, in previously untreated metastatic colorectal cancer, double-blind randomised phase 3, 151 US centres, n=455, median follow-up 20 months. Median progression-free survival was 11.8 versus 10.3 months (hazard ratio 0.92, 95% CI 0.73 to 1.16, one-sided log-rank P=.25). Overall survival was 25.6 versus 27.0 months (hazard ratio 1.05) and objective response 51 versus 44 percent (P=.12). Grade 3 or higher toxicity was not meaningfully different, and vitamin D-associated toxicities were rare. Adherence was 96 percent in both arms and the high-dose regimen reliably corrected insufficiency. A prespecified subgroup showed benefit in left-sided primaries (13.8 versus 10.2 months, hazard ratio 0.74, interaction P=.02) and harm in right-sided (hazard ratio 1.35). The authors call this exploratory and it should be read that way. The question is closed for unselected patients. The left-sided subgroup is the third time this signal has appeared, which makes it worth a dedicated trial and not worth a prescription.
Osimertinib plus chemotherapy in TP53 co-mutated EGFR-mutated NSCLC (NCT04695925). Osimertinib 80 mg daily with pemetrexed 500 mg/m2 and carboplatin AUC 5 every 3 weeks for four cycles then maintenance osimertinib plus pemetrexed, versus osimertinib alone, in treatment-naive stage IV or recurrent nonsquamous NSCLC harbouring both an EGFR-sensitising mutation and a concurrent TP53 mutation, randomised, open-label, phase 3, 17 sites in China, n=294, median follow-up 25.1 and 26.1 months. Median progression-free survival was 34.0 versus 15.6 months (difference 18.4 months, 95% CI 9.9 to 22.3; hazard ratio 0.44, 95% CI 0.32 to 0.60, P<.001), with 24-month rates of 60.8 versus 26.7 percent and median duration of response 32.7 versus 15.3 months. Benefit held across prespecified subgroups including brain metastases and L858R. Grade 3 or higher treatment-related events were 62.4 versus 14.9 percent, driven by cytopenias; 26.2 versus 1.4 percent discontinued for toxicity, and one patient died of treatment-related thrombocytopenia and pulmonary haemorrhage. Overall survival is immature at 30.6 percent (48.4 versus 36.5 months, hazard ratio 0.57, nominal P=.01). Global health status declined faster in the combination arm during induction. Single country, open-label, investigator-assessed endpoint, AstraZeneca funded. An effect this large on a prospectively selected population is hard to explain away, and it answers the question FLAURA2 left open: not whether to intensify, but in whom. TP53 status is already on most next-generation sequencing reports and most of us have been ignoring it.
SKYSCRAPER-03. Tiragolumab 840 mg plus atezolizumab 1680 mg every 4 weeks versus durvalumab, for 13 cycles as consolidation in locally advanced unresectable stage III NSCLC without progression after platinum-based concurrent chemoradiation, open-label randomised phase 3, 177 sites, n=829 (419 PD-L1 positive), median follow-up 33.0 months. The primary endpoint of independently reviewed progression-free survival in PD-L1-positive disease was not met: 19.4 versus 16.6 months, hazard ratio 0.96 (95% CI 0.75 to 1.23, P=0.76). In PD-L1 all-comers, 14.2 versus 13.8 months (hazard ratio 1.00). Overall survival was flat in both populations (PD-L1 positive hazard ratio 0.99; all-comers 45.6 versus 45.8 months, hazard ratio 0.98). Grade 3 or 4 events were similar (26.5 versus 25.7 percent), but immune-mediated events were higher with the doublet (76.4 versus 62.2 percent), as was the need for systemic corticosteroids (33.7 versus 23.5 percent) and withdrawal for toxicity (15.5 versus 9.0 percent). Grade 5 treatment-related events were lower (0.5 versus 1.7 percent). Durvalumab remains the standard after chemoradiation. Together with KEYVIBE-006 this closes anti-TIGIT in stage III, and the tiragolumab programme has already been discontinued. Worth reading as the fourth or fifth consecutive failure to improve on PACIFIC, which is itself the finding.
One issue a week. Five minutes. Every number sourced. Read by oncologists who would rather treat patients than chase PubMed.
Subscribe free →