Abstract
Background/Objectives: Colorectal cancer (CRC) is molecularly heterogeneous, and the KRAS (Kirsten rat sarcoma viral oncogene homolog) genotype both governs eligibility for anti-EGFR (epidermal growth factor receptor) therapy and carries prognostic weight. Central Asian data are scarce: no Kazakhstani series has described the KRAS variant spectrum, linked it to survival, or reported mutation detection rates across specimen types and PCR platforms. Methods: We studied 332 patients with morphologically confirmed stage II–IV colorectal adenocarcinoma. KRAS status was determined in formalin-fixed paraffin-embedded (FFPE) tumor tissue by allele-specific real-time PCR (RT-PCR) covering six selected codon 12 and 13 variants. Overall survival (OS) was estimated by the Kaplan–Meier method for the whole cohort and with stage stratification. Separately, mutation detection rates were recorded in three non-overlapping groups of different patients: plasma RT-PCR on the Idylla platform (n = 30), plasma nanoplate digital PCR (dPCR) on QIAcuity One (n = 120), and a routine tissue RT-PCR series (546 evaluable of 550). Because these groups differed in patients, specimen type, and mutation panel, this comparison describes observed detection rates only and supports no inference about analytical sensitivity, specificity, or concordance. Results: KRAS was mutated in 149/332 tumors (44.9%); codon 12 supplied 80.5% of variants, led by G12D (32.2%), G12V (24.8%) and G13D (19.5%). Median OS did not differ between mutant and wild-type tumors (39.0 vs. 36.6 months; p = 0.40). Variant-level differences were directionally consistent, but none was significant, and all were exploratory and unadjusted for multiplicity. Observed detection rates were 40.3% for tissue RT-PCR (220/546), 13.3% for plasma RT-PCR (4/30; continuity-corrected p = 0.006 vs. tissue), and 50.8% for plasma dPCR (61/120; continuity-corrected p = 0.044, Pearson p = 0.034 vs. tissue). BRAF V600E was detected by plasma dPCR in 11/120 cases (9.1%). Conclusions: This first Kazakhstani series places KRAS frequency within the internationally reported range and shows that variant-level reporting reveals prognostic structure that a binary call conceals. The higher detection rate seen with plasma dPCR is hypothesis-generating, not evidence of platform superiority, and motivates a prospective paired-sample study with harmonized mutation panels.