Total Neoadjuvant Therapy (TNT)

Strategy for locally advanced Rectal Cancer: deliver all planned chemotherapy and radiotherapy before surgery (rather than the traditional split of neoadjuvant chemoradiotherapy → surgery → adjuvant chemotherapy). Goals: tumor downstaging, eradication of distant micrometastases, and — increasingly — organ preservation via non-operative management (NOM) for strong responders.

Originally built from a single secondary review; now cross-validated against the primary publications of all four headline phase III TNT trials — Bahadoer 2021 - RAPIDO Trial, Conroy 2024 - PRODIGE 23 Long-Term Results, Jin 2022 - STELLAR Trial, and Ciseł 2019 - POLISH II Long-Term Results — plus Schrag 2023 - PROSPECT Trial, a related but distinct radiation-omission trial in a lower-risk population, and Dijkstra 2023 - RAPIDO 5-Year Locoregional Failure, RAPIDO’s extended follow-up. The TNT trials disagree with each other on the locoregional-recurrence trade-off (see Efficacy below): RAPIDO now shows a confirmed, statistically significant local-control cost at 5 years, PRODIGE 23 and STELLAR show numerically better local control, and POLISH II shows no difference in either direction — flagged explicitly as a real, RAPIDO-specific finding rather than a class effect.

Why TNT over standard neoadjuvant CRT + adjuvant chemo

• Compliance is much better neoadjuvantly: 80-90% of patients complete planned chemo before surgery, vs. only 25-50% completing adjuvant chemo after (perioperative complications, stoma, deconditioning all interfere with adjuvant dosing). This is possibly the single most practical argument for TNT, independent of the survival numbers.
• Shortens total oncologic treatment time by ~1-2 months.
• Early systemic therapy start plausibly better addresses micrometastatic disease that’s already present at diagnosis (most distant relapses manifest within the first 2-5 years).

Efficacy

Now grounded in four trials’ primary publications, not just the secondary review’s Table 1, plus RAPIDO’s own 5-year extended follow-up:

Trial	Design	n (TNT/ctrl)	pCR (TNT vs ctrl)	3y-DFS	OS	Locoregional recurrence
Bahadoer 2021 - RAPIDO Trial (3y) / Dijkstra 2023 - RAPIDO 5-Year Locoregional Failure (5.6y)	SCRT → 6xCAPOX/9xFOLFOX4 → TME, vs CRT → TME ± adjuvant	462/450 (460/446 in 5y LRR analysis)	28% vs 14%	(3y-DrTF 23.7% vs 30.4%; 5y-DrTF 27.8% vs 34.0%, HR 0.79, P=0.048)	3y-OS 89.1% vs 88.8% (NS); 5y-OS 81.7% vs 80.2% (HR 0.91, P=0.50, NS)	LRR after R0/R1 resection: 10.2% vs 6.1% at 5.6y, P=0.027 — significant, TNT worse (not significant at the original 3y analysis)
Conroy 2024 - PRODIGE 23 Long-Term Results	6x mFOLFIRINOX → CRT → TME → adjuvant, vs CRT → TME → adjuvant	231/230	28% vs 14%¹	7y-DFS 67.6% vs 62.5%	7y-OS 81.9% vs 76.1% (p=.033)	5.3% vs 8.1% at 7y — TNT numerically better
Jin 2022 - STELLAR Trial	SCRT → 4xCAPOX → TME → 2xCAPOX, vs CRT → TME → 6xCAPOX	302/297	21.8% vs 12.3%²	3y-DFS 64.5% vs 62.3% (noninferior)	3y-OS 86.5% vs 75.1% (secondary, non-powered)	8.4% vs 11.0% at 3y — TNT numerically better
Ciseł 2019 - POLISH II Long-Term Results	SCRT → 3xFOLFOX4 → TME, vs CRT (50.4Gy+oxaliplatin) → TME — higher-risk cT4/fixed-cT3 population	261/254	16% vs 12% (ypT0N0, NS)	8y-DFS 43% vs 41% (HR 0.95, NS)	8y-OS 49% vs 49% (HR 0.90, NS) — early 3y OS edge for TNT (9%) did not persist	35% vs 32% cumulative local failure at 8y (HR 1.08, NS) — no difference either direction

¹ PRODIGE 23’s pCR figure is from the original 3-year primary publication (Conroy 2021, not independently ingested), reproduced in the long-term follow-up paper’s discussion; not itself re-reported as a headline number in the long-term paper. ² Figure is pCR + sustained clinical complete response combined, not pCR alone.

• DFS/MFS/distant-metastasis reduction is consistent across the three trials that showed a benefit (RAPIDO, PRODIGE 23, STELLAR) — TNT reliably reduces distant recurrence risk vs standard CRT in those. POLISH II, testing a similar SCRT+consolidation-chemo design in a higher-risk cT4/ fixed-cT3 population, showed no DFS or distant-metastasis benefit — a genuine outlier worth keeping in mind before treating “TNT reduces distant recurrence” as universal across all populations and regimens.
• OS benefit is trial-dependent, not universal. PRODIGE 23 (chemotherapy-first design, same CRT in both arms) shows a clear, mature 7-year OS benefit. STELLAR shows a numerically favorable but statistically non-powered OS trend at only 35 months’ follow-up. RAPIDO (SCRT-substitution design) shows no OS benefit at 3 years, which its own discussion attributes partly to an “ATRESS phenomenon” (shorter survival after metastatic relapse in the TNT arm) — a phenomenon PRODIGE 23’s authors explicitly checked for and did not find in their own data — and this remains true at 5.6 years’ follow-up (HR 0.91, P=0.50). POLISH II shows no OS difference at 8 years, after an early 3-year advantage for the short-course/consolidation-chemo arm that did not persist — a useful cautionary example of an early survival signal that later washed out with longer follow-up.
• The locoregional-recurrence trade-off is now confirmed as a real, RAPIDO-specific finding — not a class effect of TNT, and not resolved to a single directional conclusion across trials. At RAPIDO’s original 3-year analysis this was a numerical, non-significant trend; at 5.6 years’ follow-up (Dijkstra 2023 - RAPIDO 5-Year Locoregional Failure) it became statistically significant (LRR 10.2% vs 6.1% after R0/R1 resection, P=0.027), with intraoperative mesorectal breach, positive CRM, tumor deposits, and higher ypN-stage as independent correlates — the authors speculate (but do not prove) that RAPIDO’s long overall treatment time (~40 weeks EXP vs ~25 weeks STD) may yield a more fragile/fibrotic mesorectal specimen. PRODIGE 23 and STELLAR’s TNT arms continue to show numerically better local control than CRT, and POLISH II shows no difference in either direction. Both STELLAR and RAPIDO use SCRT-based TNT regimens, so the trade-off isn’t simply explained by SCRT vs chemotherapy-first design. Do not treat “TNT trades distant control for local control” as an established generalization across the class — it is now a well-confirmed finding specific to RAPIDO’s regimen/population/treatment-time profile, not observed in PRODIGE 23, STELLAR, or POLISH II.
• pCR (pathologic complete response) roughly doubled vs CRT alone in RAPIDO, PRODIGE 23, and STELLAR, to ~22-28% in the TNT arms — the key enabler of NOM, though pCR rate and successful long-term NOM are not tightly correlated (see NOM section). POLISH II’s ypT0N0 rate (16% vs 12%) was not significantly different between arms.
• Toxicity: RAPIDO and STELLAR both showed higher acute grade ≥3 toxicity with TNT during the preoperative phase (RAPIDO 48% vs 25%; STELLAR 26.5% vs 12.6%, mostly hematologic) — real practical cost not fully captured by the efficacy headline numbers. Treatment-related death rates were low and similar between arms across trials.
• A related but distinct trial, Schrag 2023 - PROSPECT Trial, tests the opposite strategy — omitting radiotherapy for good responders in a lower-risk, sphincter-sparing-eligible population (not a TNT trial in this concept’s sense) — see its source page and the Indications table below.

Indications — risk-stratified, not blanket

Tumor profile	TNT indicated?
T4, N2, MRF+ (mesorectal fascia involved), or EMVI+ (extramural vascular invasion)	Yes — clearest indication
T3N1	Consider, based on other risk factors
T3N0, upper/mid rectum, favorable histology	Usually no benefit — standard (chemo)RT or nothing may suffice
T3N0, distal rectum	TNT worth considering specifically for organ-preservation/NOM potential — highest pCR rates, biggest QoL payoff (avoids permanent colostomy)
Stage I	Investigational only — unclear benefit over standard CRT; NOM burden may not be justified if surgery alone would cure without permanent colostomy
Oligometastatic (stage IV, resectable mets)	Reasonable extrapolation, but evidence is sparse — no dedicated large trials yet
MMRd/MSI tumors (~5% of rectal cancers)	Different pathway entirely — immunotherapy (PD-1/PD-L1 inhibitors: ipilimumab+nivolumab, dostarlimab, etc.) achieves 60-100% pCR, even without radiotherapy. Chemo-based TNT not recommended for this subgroup. MMRd/MSI testing should be done before neoadjuvant therapy decisions.

Modalities

Radiotherapy — two regimens, no proven superiority of one over the other in the TNT setting:

• CRT (long-course chemoradiotherapy): 45Gy/25fx to tumor+mesorectum+nodes over 5 weeks, boost to 50Gy, + concurrent capecitabine. Better-predictable late toxicity (normofractionation); preferred for very advanced tumors (T4/MRF+), borderline-performance-status patients, and NOM candidates.
• SCRT (short-course): 25Gy/5fx over 1 week. Lower total equivalent dose; convenient, doesn’t delay systemic chemo; may be preferred when high risk of/present metastatic disease. Possible (unproven) higher late local relapse rate vs. CRT.

Chemotherapy — CAPOX or FOLFOX6 standard (3-4 months = optimal TNT length; no evidence longer helps). FOLFOXIRI/FOLFIRINOX (add irinotecan) is more intensive, reasonable for very-high-risk/ oligometastatic disease as conversion therapy, but hasn’t proven superior to CAPOX/FOLFOX by direct randomized comparison — Alvarez 2024 - JANUS Trial Protocol (Alliance A022104/NRG-GI010) is testing triplet (mFOLFIRINOX) vs doublet (mFOLFOX6/CAPOX) specifically as consolidation chemo after long-course CRT, powered for clinical complete response (phase II) and DFS (phase III); still accruing (330/760 as of May 2024), no results yet. Separately, Brouquet 2020 - GRECCAR16 Trial Protocol (NORAD01-GRECCAR16) is testing modified FOLFIRINOX as a replacement for radiotherapy entirely (not consolidation) in CRM-negative resectable LARC — also no results yet. Oxaliplatin-based regimens carry more toxicity in patients >70-75yo or with major comorbidities — individualize.

Targeted therapy (bevacizumab, cetuximab/panitumumab) — not currently recommended outside trials; results equivocal, increased toxicity (bevacizumab: healing problems, fistula risk — must be held 4-6 weeks before surgery and around radiotherapy).

Immunotherapy — first-line for MMRd/MSI tumors (see indications table above); optimal duration unclear (2 weeks-6 months across studies).

Sequence and timing

• No clearly defined optimal sequence; radiotherapy-first may give more time for local response to manifest and is favored for very advanced/bleeding tumors or NOM intent. Chemotherapy-first may suit high distant-metastatic-risk patients (immediate systemic control).
• Gap between modalities kept short (~10-14 days) to allow acute-toxicity recovery without letting resistant clones regrow.
• Surgery typically 3-6 weeks after chemo, 4-10 weeks after radiotherapy (or immediately after SCRT). Evidence for optimal interval is weak; longer intervals allow fuller response but also more time for regrowth of resistant cells — if a longer interval is chosen, re-stage before surgery.

Surgery and non-operative management (NOM)

• Standard surgery: low anterior resection (LAR) or abdominoperineal resection (APR), with total mesorectal excision (TME). No significant difference in R0 rate (~90%+), Clavien-Dindo complications, or LARS between TNT and standard arms.
• NOM (“watch and wait”): offered to patients with complete/near-complete clinical response, assessed 4-8 weeks post-TNT by DRE + endoscopy + MRI (MSK three-tiered scheme). Requires a highly competent multidisciplinary team and a well-informed, compliant patient willing to undergo intensive surveillance. Repeat biopsy after TNT is not recommended (false pos/neg risk, healing complications).
  • Locoregional relapse on surveillance: ~20-40%, most salvageable by radical surgery without DFS penalty.
  • Distant metastatic risk while on surveillance: 5-15%, mostly associated with locoregional recurrence — this is the main real risk of NOM and should be discussed candidly with patients.
  • pCR rate and NOM success aren’t tightly correlated — OPRA trial’s 3-year TME-free survival (41-50%, depending on sequence) exceeded PRODIGE 23’s pCR rate (28%) by >10 points, suggesting some patients achieve complete response later than the standard pCR-assessment timepoint.

Adjuvant therapy after TNT

Not indicated if the full neoadjuvant course (3-4 months, ideally 4) was completed — splitting chemo pre/post-op with a perioperative gap plausibly lowers total dose intensity and may promote resistance. Exceptions: R1-2 resections (3 months adjuvant reasonable), or patients who didn’t complete ≥3 months of neoadjuvant chemo (adjuvant capecitabine ×3mo as a fallback).

Risk-based approach (summary of the review’s Fig. 1 algorithm)

Stratifies by age (>40 vs <40), tumor location, histologic grade, MSS/MMRp vs MSI/MMRd, T-stage, and N-stage into four broad pathways:

1. Locally advanced, obstructing/bleeding → CRT or SCRT → CAPOX/FOLFOX6 → surgery
2. Locally advanced + high risk of distant mets → SCRT → CAPOX/FOLFOX6 (or interleaved)
3. High risk of / manifest distant mets → CAPOX/FOLFOX6/FOLFOXIRI → SCRT (chemo-first)
4. MSI/MMRd → SCRT + ICI, or ICI alone

Open items / gaps

• RAPIDO, PRODIGE 23, STELLAR, and POLISH II primary publications are now all ingested (2026-08-20), along with RAPIDO’s 5-year extended follow-up. The comparison table above now has four independent trials.
• The locoregional-recurrence trade-off question (previously flagged unresolved) is now answered for RAPIDO specifically: Dijkstra 2023 - RAPIDO 5-Year Locoregional Failure confirms a statistically significant increase in LRR with TNT at 5.6 years. It remains a RAPIDO-specific finding, not observed in PRODIGE 23, STELLAR, or POLISH II — see Efficacy above for the full synthesis. This should be revisited only if further comparative literature (e.g. a formal meta-analysis across all four trials) is ingested.
• Rectal Cancer is no longer a stub — since expanded with staging, TME/CRM/downstaging history, and survival by stage (via the Sabiston Rectal Cancer Surgery chunk) — but MRF and EMVI specifically, referenced here as TNT risk-stratification criteria, still aren’t independently defined with their own detail anywhere in the wiki.
• JANUS and NORAD01-GRECCAR16 protocols are now ingested (Alvarez 2024 - JANUS Trial Protocol, Brouquet 2020 - GRECCAR16 Trial Protocol) — both are still accruing/awaiting results, so the substantive triplet-vs-doublet-consolidation and chemotherapy-vs-radiotherapy-omission questions remain open. ACO/ARO/AIO-18 (CRT vs. SCRT head-to-head) and GRECCAR 14 are still referenced only secondhand (via the Boublikova review and PRODIGE 23’s long-term discussion, respectively) — worth ingesting their protocols or results if/when available.
• Schrag 2023 - PROSPECT Trial (radiation-omission in lower-risk LARC) is now ingested and referenced above — worth a fuller writeup on Rectal Cancer’s treatment-selection section if that page is expanded further to cover treatment individualization by risk group.