Genomic Scars May Predict Which Rectal Cancers Respond to Radioimmunotherapy

Rectal cancer remains one of the world’s most formidable malignancies, accounting for roughly one in ten of all cancer diagnoses and standing as the second leading cause of cancer-related death according to the 2022 GLOBOCAN statistics. For patients whose tumors have grown into surrounding tissue or spread to nearby lymph nodes, standard care has long relied on a punishing sequence: preoperative chemoradiotherapy, radical surgery to remove the rectum, and then months of adjuvant chemotherapy. Yet this multimodal strategy achieves a pathological complete response, meaning no residual tumor cells are found in the surgical specimen, in only about 10 to 30 percent of patients. That gap has pushed oncologists toward immunotherapy, which has produced spectacular results in the minority of colorectal cancers with deficient DNA mismatch repair, but has largely failed in the roughly 85 percent of patients whose tumors are microsatellite stable and immunologically cold.

A new study published in Cancer Reports by researchers at the Third Affiliated Hospital of Army Medical University in China now offers a provocative twist to that story. The team investigated whether a genomic measure called the homologous recombination deficiency score, or HRD score, could predict which patients with mismatch repair proficient, locally advanced rectal cancer would respond to a novel neoadjuvant regimen combining short-course radiotherapy with sequential chemotherapy and the immunotherapy drug tislelizumab. What they found challenges a long-standing assumption in oncology: the relationship between HRD and treatment response, at least in this setting, is not a simple case of higher being better.

Homologous recombination is one of the cell’s most elegant repair systems, faithfully mending dangerous double-strand breaks in DNA. When this pathway fails, a condition known as homologous recombination deficiency, the genome accumulates characteristic scars: loss of heterozygosity across large chromosomal stretches, telomeric allelic imbalance, and large-scale state transitions. These three metrics can be combined into a single HRD score, which has already proven clinically valuable. In triple-negative breast cancer, for example, patients with high HRD scores respond markedly better to platinum chemotherapy, and tumors with BRCA1 or BRCA2 mutations, which cripple the homologous recombination machinery, are famously sensitive to both platinum drugs and radiation. Because radiation therapy kills cells largely by inducing exactly the kind of DNA double-strand breaks that homologous recombination repairs, the logic of using HRD to predict radiosensitivity is compelling.

The Chinese team enrolled 32 patients with mismatch repair proficient locally advanced rectal cancer in a clinical trial running from 2021 to 2023. Before any treatment began, colonoscopy biopsies were taken and subjected to whole-exome sequencing, bulk RNA sequencing, and a targeted HRD scoring assay based on roughly 40,000 single nucleotide polymorphism probes across the genome. Patients then received short-course radiotherapy followed by capecitabine and oxaliplatin chemotherapy together with tislelizumab, an antibody that blocks the PD-1 immune checkpoint. The results of the regimen itself were striking: 59.38 percent of patients achieved a clinical complete response confirmed by endoscopy, and 40.62 percent reached a pathological complete response, a rate well above what conventional chemoradiotherapy typically delivers.

But when the researchers correlated the HRD scores with pathological outcomes, the expected pattern failed to materialize. Across the entire cohort, there was no significant difference in HRD score between patients who achieved complete pathological remission and those who did not. The two patients with the highest scores in the cohort, 70 and 67, both achieved complete responses, but the bulk of the data told a stranger story. After excluding those two outliers and reanalyzing the remaining 30 cases, the difference became borderline significant, and only the HRD score, not tumor mutation burden or mutations in homologous recombination genes, showed statistically meaningful power to distinguish responders from non-responders, with an area under the receiver operating characteristic curve of 0.710.

The most intriguing finding emerged when the team divided patients into three groups by HRD score: low, from 0 to 8; intermediate, from 9 to 37; and high, above 37. The pathological complete response rate was 57.1 percent in the low group but only 18.8 percent in the intermediate group, a difference that reached statistical significance. Not a single patient in the high-score group failed to respond, though that group was small. This non-monotonic pattern, in which the middle of the distribution fares worst, upends the conventional wisdom that greater genomic instability should translate uniformly into greater treatment sensitivity.

To understand why, the researchers turned to the tumor immune microenvironment. Using computational tools that infer immune cell abundance from gene expression data, they found that the abundance of CD8+ T cells, the cytotoxic lymphocytes that immunotherapy depends upon, declined as HRD score rose from low to intermediate, and myeloid dendritic cells, the antigen-presenting sentinels that prime T cell responses, decreased sequentially across all three groups. Gene set enrichment analysis revealed that tumors in the low-score group were significantly enriched in pathways governing dendritic cell maturation, MHC protein complexes, T cell proliferation, and hypoxia compared with the intermediate group. Meanwhile, tumor mutation burden actually increased with HRD score, meaning the intermediate tumors carried more mutations yet presented them less effectively to the immune system.

The authors propose a hypothesis rooted in the biology of copy number variations. Microsatellite stable colorectal cancers typically carry near-diploid genomes riddled with copy number alterations, and prior research has shown that high degrees of such alterations correlate with reduced CD8+ T cell infiltration and immune evasion. Since HRD is largely driven by copy number changes, the team suggests that in the low-to-intermediate range, rising HRD may reflect mounting copy number burden that progressively suppresses the immune microenvironment, leaving too few pre-existing immune cells for short-course radiotherapy to mobilize. Radiation-induced immunogenic cell death works best when the cGAS-STING pathway and an existing immune infiltrate are available to amplify the signal; a barren microenvironment may simply have nothing to amplify. In contrast, tumors with the highest HRD scores may benefit from a dominant effect of intrinsic radiosensitivity, reflected in a significantly lower radiosensitivity index in the high-score group, which overwhelms any immune suppression.

The study’s authors are careful to acknowledge its limitations. Thirty-two patients is a small sample, too small for multivariate regression to identify independent predictors of response, and the HRD scoring method used was not benchmarked against clinically approved next-generation sequencing platforms, so no clinically meaningful cutoff value could be established. The observed associations also cannot be attributed to radiotherapy alone, since every patient also received chemotherapy and immunotherapy. Still, the implications are considerable. If validated in larger cohorts, HRD scoring could help oncologists identify which mismatch repair proficient rectal cancer patients are likely to benefit from short-course radiotherapy combined with immunotherapy, and which might need entirely different strategies, such as interventions designed to inflame a cold tumor before radiation begins.

For a disease that affects millions worldwide and for the vast majority of colorectal cancer patients who have been excluded from the immunotherapy revolution, the message from this study is one of cautious optimism tempered by biological complexity. Genomic scars, it appears, do not merely record a tumor’s history of DNA repair failure; they may also shape the immunological battlefield on which modern multimodal therapy must fight. Understanding why intermediate HRD tumors resist a regimen that succeeds elsewhere in the HRD spectrum could unlock the next generation of biomarker-guided, immune-activating treatment for one of oncology’s most stubborn cold tumor populations.

Subject of Research: HRD score as a predictive biomarker for neoadjuvant short-course radiotherapy and chemoimmunotherapy in pMMR locally advanced rectal cancer

Article Title: Association of HRD Score With Efficacy of Neoadjuvant Short‐Course Radiotherapy Plus Sequential Chemoimmunotherapy in pMMR Locally Advanced Rectal Cancer

Article References: Ran, X., Xiao, H., Zhou, P., He, J., Si, M., Chen, X., Su, X., Chen, Z., Du, J., Dai, X., Yang, X., Shen, H., He, Z., Li, F., Li, M., & Chen, C. (2026). Association of HRD Score With Efficacy of Neoadjuvant Short‐Course Radiotherapy Plus Sequential Chemoimmunotherapy in pMMR Locally Advanced Rectal Cancer. Cancer Reports, 9(10), Article e70708. https://doi.org/10.1002/cnr2.70708

Image Credits: AI Generated

DOI: 10.1002/cnr2.70708

Keywords: rectal cancer, HRD score, homologous recombination deficiency, short-course radiotherapy, tislelizumab, immunotherapy, pMMR, pathological complete response, tumor microenvironment, copy number variation, biomarker, neoadjuvant therapy