Why this matters
Immunotherapy has transformed the management of mismatch repair-deficient rectal cancer, but the overwhelming majority of rectal cancers are proficient mismatch repair, where checkpoint inhibitors alone have limited activity.
An important unanswered question is whether radiotherapy and chemotherapy can create a sufficiently immunogenic environment for PD-1 blockade to improve response in pMMR disease.
STELLAR II directly tested that strategy in a large randomized trial.
The short-term result is striking: adding sintilimab to short-course radiotherapy-based total neoadjuvant therapy nearly doubled the overall complete response rate.
Whether that translates into better long-term disease control remains unknown.
Study design
STELLAR II is a multicenter randomized phase 2/3 trial in patients with mid-to-low, proficient mismatch repair locally advanced rectal cancer.
Patients were randomized to:
- short-course radiotherapy followed by CAPOX
- the same short-course radiotherapy and CAPOX backbone plus sintilimab
Short-course radiotherapy was:
- 5 Gy x 5 fractions
Both groups then received 4 cycles of CAPOX.
Patients in the experimental group also received sintilimab during neoadjuvant treatment.
Following response assessment, patients proceeded to either:
- surgical resection
- watch-and-wait after clinical complete response
Two additional cycles of the previously assigned systemic regimen were subsequently planned.
The phase II primary endpoint was complete response.
The phase III primary endpoint is 3-year disease-free survival.
A total of 588 patients were randomized:
- 296 to immunotherapy-containing TNT
- 292 to standard TNT
The population was clinically high risk:
- 94.9% had T3-4 disease
- 92.9% were node positive
- 71.3% had lower rectal tumors
- 48.8% had threatened or involved mesorectal fascia
- 45.4% had extramural venous invasion
All patients completed short-course radiotherapy.
Key results
The overall complete response rate was substantially higher with sintilimab.
Complete response occurred in:
- 51.9% with immunotherapy-containing TNT
- 26.8% with standard TNT
- p<0.001
This represents an absolute increase of approximately 25 percentage points.
Among patients who underwent surgery, pathological complete response was:
- 41.7% with sintilimab
- 19.4% without sintilimab
The pathological response rate therefore more than doubled.
A total of 92 patients, representing 15.6% of the study population, entered a watch-and-wait strategy after achieving clinical complete response.
Treatment completion
Treatment delivery remained high in both groups.
Completion of all 4 planned chemotherapy cycles was:
- 92.9% with immunotherapy-containing TNT
- 94.1% with standard TNT
In the experimental arm, 86.1% of patients received at least 4 cycles of sintilimab.
All randomized patients completed short-course radiotherapy.
Toxicity
The improvement in response came with additional toxicity.
Grade 3-4 adverse events occurred in:
- 27.7% with sintilimab
- 19.9% with standard TNT
Grade 3-5 immune-related adverse events occurred in:
- 3.4% of patients receiving immunotherapy
The approximately 8 percentage point increase in severe toxicity is therefore an important part of the treatment tradeoff.
Interpretation
STELLAR II provides one of the strongest randomized signals so far that immune checkpoint blockade can meaningfully increase response to neoadjuvant treatment even in pMMR rectal cancer.
The size of the effect is difficult to dismiss.
Overall complete response increased from 26.8% to 51.9%, while pathological complete response increased from 19.4% to 41.7%.
For rectal cancer, that matters for more than pathological staging.
A higher complete response rate potentially increases the number of patients who can pursue organ preservation and avoid total mesorectal excision.
That makes STELLAR II particularly relevant in the current era of watch-and-wait strategies.
But complete response is not the same as long-term oncologic benefit.
The phase III primary endpoint is 3-year disease-free survival, and those data are not yet mature.
This distinction is crucial because a treatment can increase pathological or clinical complete response without necessarily improving metastasis control, disease-free survival, or overall survival.
The toxicity tradeoff also matters. Grade 3-4 adverse events increased from 19.9% to 27.7% with the addition of sintilimab.
The current data therefore support a strong biological and clinical activity signal, particularly for organ-preservation strategies, but not yet a definitive new standard for all patients with pMMR locally advanced rectal cancer.
Limitations
The current analysis reports short-term efficacy rather than the definitive phase III endpoint.
Three-year disease-free survival remains pending.
The complete response endpoint combines pathological complete response after surgery with clinical complete response in patients managed with watch-and-wait, which reflects clinically relevant treatment pathways but introduces different methods of response confirmation.
Long-term local regrowth after watch-and-wait, distant metastasis, salvage treatment, and overall survival remain unknown.
The additional toxicity associated with sintilimab must also be weighed against the potential organ-preservation benefit.