Schrag 2023 — PROSPECT Trial

Citation: Schrag D, Shi Q, Weiser MR, et al. Preoperative Treatment of Locally Advanced Rectal Cancer. N Engl J Med. 2023;389(4):322-334. doi:10.1056/NEJMoa2303269.

Filename note: Stephen’s uploaded file was named NEJMoa2303269_RAPIDO.pdf. Despite the “RAPIDO” in the filename, this is not the RAPIDO trial — the DOI (10.1056/NEJMoa2303269) and full text confirm it is the PROSPECT trial (Schrag et al., NEJM 2023). The real RAPIDO trial (Bahadoer et al., Lancet Oncol 2021) was separately uploaded and is filed as Bahadoer 2021 - RAPIDO Trial. Flagging this explicitly since the mislabel could otherwise cause real confusion later.

Study design

• Alliance for Clinical Trials in Oncology, multicentre (264 institutions, Canada/Switzerland/USA), randomised, unblinded, noninferiority, seamless phase 2–3 trial (NCT01515787). 1194 patients randomised June 2012–December 2018 (1128 began treatment: 585 FOLFOX / 543 chemoradiotherapy).
• Population: rectal adenocarcinoma clinically staged cT2 node-positive, cT3 node-negative, or cT3 node-positive, who were candidates for sphincter-sparing surgery. Key exclusions: T4 tumours, ≥4 pelvic lymph nodes with short axis >10mm, or tumour within 3mm of the radial margin. This is a materially lower-risk population than RAPIDO, PRODIGE 23, or STELLAR — those trials targeted higher-risk, higher-stage disease.
• Experimental arm: neoadjuvant FOLFOX (6 cycles modified FOLFOX6), with chemoradiotherapy reserved only for poor responders (primary tumour shrinkage <20%, or FOLFOX discontinued for toxicity) → surgery. Postoperative adjuvant FOLFOX suggested (not mandated) if resection not pathologically complete.
• Control arm: standard neoadjuvant chemoradiotherapy (50.4 Gy/28 fractions + sensitizing fluoropyrimidine) → surgery → postoperative adjuvant FOLFOX suggested (not mandated).
• Primary endpoint: DFS noninferiority (prespecified margin: HR<1.29, corresponding to no more than a 5-percentage-point absolute reduction in 5-year DFS).
• Median follow-up 58 months.

Key findings

Outcome (5-year)	FOLFOX (selective CRT)	Chemoradiotherapy	HR (95% CI)	p
Disease-free survival	80.8%	78.6%	0.92 (90.2% CI 0.74–1.14)	.005 for noninferiority
Overall survival	89.5%	90.2%	1.04 (0.74–1.44)	—
Local recurrence	1.8%	1.6%	1.18 (0.44–3.16)	—
Pathological complete response	21.9%	24.3%	—	—

• Noninferiority for DFS confirmed — neoadjuvant FOLFOX with selective (not routine) chemoradio- therapy was noninferior to routine chemoradiotherapy.
• 89.6% of the FOLFOX group avoided chemoradiotherapy entirely — this is the central practical finding: most patients in this lower-risk, sphincter-sparing-eligible population can skip pelvic radiation without compromising oncologic outcomes.
• Local recurrence was very low and similar in both groups (<2% at 5 years) — unlike RAPIDO’s locoregional-control question, radiation omission here did not measurably compromise local control in this lower-risk population.
• Toxicity trade-off by phase of treatment: severe (grade ≥3) adverse events were higher with FOLFOX during the neoadjuvant phase (41.0% vs 22.8%) — but the FOLFOX treatment period was roughly twice as long (12 weeks vs 5.5 weeks) as chemoradiotherapy, so this isn’t adjusted for treatment duration. Postoperatively, severe adverse events were lower with FOLFOX (25.6% vs 32.6%).

Limitations

• Explicitly not a TNT trial in the RAPIDO/PRODIGE-23/STELLAR sense — PROSPECT tests whether radiotherapy can be omitted for good responders in a lower-risk, sphincter-sparing-eligible population, not whether treatment intensity/sequencing should be increased for higher-risk disease. The two questions (de-intensification vs TNT-style intensification) address different clinical scenarios and shouldn’t be conflated when citing “TNT trial” evidence.
• Excludes T4 tumours, bulky nodal disease, and close-margin disease — the population most likely to benefit from radiotherapy is excluded by design, so these results cannot be extrapolated to higher-risk LARC.
• Unblinded (open-label) — inherent to the intervention (radiation vs no radiation is not blindable).
• 90.2% (not 95%) CI used for the primary noninferiority comparison — a wider, less conservative interval than the 95% CI convention, per the trial’s noninferiority design; worth noting when comparing precision against superiority trials like RAPIDO/PRODIGE 23.

Relevance

Not one of the primary trials originally sought for Total Neoadjuvant Therapy (TNT)‘s evidence gap, but a valuable complementary/contrasting trial uploaded alongside them (with the filename mislabel noted above). Where RAPIDO, PRODIGE 23, and STELLAR test intensifying and resequencing neoadjuvant treatment for higher-risk LARC, PROSPECT tests de-intensifying (radiation-omission) treatment for lower-risk, sphincter-sparing-eligible LARC. Both are part of the broader modern rectal-cancer treatment-individualization story and belong in Rectal Cancer’s treatment-selection discussion, but should be filed as answering a different clinical question than the core TNT trials rather than merged into one “TNT evidence” bucket.