Citations

Background Tumors with mismatch repair deficiency (MMRd) classically display concomitant loss of MLH1/PMS2 or MSH2/MSH6 on immunohistochemistry (IHC) with microsatellite instability-high (MSI-High) status on molecular testing. Nevertheless, a different phenotype can occur in up to 15% of MMRd tumors (unusual phenotype). Data on the efficacy of immunotherapy in this population remain scarce. Methods We conducted a retrospective study within the IMMUNODIG MSI cohort including patients with advanced gastrointestinal MMRd tumors treated with immune checkpoint blockade in the real-world setting. We selected patients with both IHC and MSI assays data available. Unusual MMRd tumors were classified into four distinct groups: (1) isolated loss of PMS2 or MSH6 with MSI-H (isolated/MSI-H), (2) complex loss of MMR proteins with MSI-H (complex/MSI-H), (3) loss of one or more MMR proteins without MSI-H (MMRd-IHC/microsatellite stability (MSS) or MSI-low (MSI-L)), and (4) four MMR proteins retained with MSI-H (retained IHC/MSI-H). Results Out of 759 patients in the IMMUNODIG-MSI cohort, 571 patients met inclusion criteria. Of these, 90 (15.8%) had an unusual phenotype (47 isolated/MSI-H, 19 complex/MSI-H, 16 MMRd-IHC/MSS or MSI-L and 8 retained IHC/MSI-H). Compared with classical phenotypes, patients with a tumor harboring an unusual phenotype had a younger age at treatment (p=0.005), increased RAS mutation (p=0.005), reduced BRAF p.V600E mutation rates (p<0.001), a higher proportion of Lynch syndrome (p<0.001) and a higher prevalence of non-colorectal cancers (p=0.021). After a median follow-up of 28.1 months (mo), there was a significant difference in progression-free survival, with median values of not reached, 66.4 mo, 37.2 mo, 18.3 mo and 5.5 mo for complex/MSI-H, isolated/MSI-H, classical, retained IHC/MSI-H and MMRd-IHC/MSS or MSI-L subgroups, respectively (p<0.001). Notably, objective response rates were 59.1%, 58.7%, and 63.2% for complex/MSI-H, isolated/MSI-H and classical contrasting with 50% and 25% for retained IHC/MSI-H and MMRd-IHC/MSS or MSI-L, with no complete response observed in the latter two. Conclusion Our findings underscore the need for dual testing and advocate for the presence of both MMRd-IHC and MSI-H for optimal immunotherapy response. Of note, complex MMR aberrations and isolated PMS2/MSH6 losses with MSI-H may represent promising candidates for enhanced immunotherapy efficacy.

PURPOSE The development of colorectal liver metastases (CRLMs) is associated with poor prognosis, and recent data suggest that metastasis to the liver is associated with resistance to immunotherapy. We characterized the microenvironment of primary colorectal carcinomas (CRCs) relative to their synchronous CRLMs using a validated segmentation algorithm (QuantCRC) that quantifies 15 distinct morphologic tumor features. MATERIALS AND METHODS Adult patients with CRC with synchronous CRLM (N = 57) at Mayo Clinic were identified from the electronic health record. Routine tumor hematoxylin and eosin sections were digitized and reviewed for quality control. QuantCRC (Aiforia, Inc) was applied to digitized images to extract 15 GI pathologist predefined morphological features. Tumor features were compared between paired primary tumors and their synchronous CRLMs using the Wilcoxon signed-rank test overall and within subgroups. Linear regression models were used to find predictors of the paired differences in the distinct tumor morphological features. RESULTS The study included 57 patients (median age 59 years [IQR, 50-73], 51% female) with CRC primary and synchronous CRLM, and among primaries, 37 (65%) were left-sided. QuantCRC identified six of 15 morphological features that differed significantly between primaries and their CRLMs. Compared with the primary, CRLM showed reduced stroma, more high-grade tumor and necrosis, higher tumor–stromal ratio (TSR; all P < .05), and a trend toward increased tumor-infiltrating lymphocyte (TIL) density ( P = .052). Synchronous CRLM from patients with left-sided primary tumors had significantly higher TSR, percent high-grade, percent necrosis, and TIL density compared with the primary (all P values ≤0.04). Among right-sided primaries, CRLM had a significantly reduced percent mature stroma ( P = .0296) and a higher percent necrosis ( P = .0027). CONCLUSION Compared with the primary, synchronous CRLMs showed a higher TSR and increased high-grade tumor and necrosis, which are features of tumor aggressiveness and a metastatic phenotype.

LBA1 Background: Standard adjuvant chemotherapy of stage III colon cancer consists of a fluoropyrimidine plus oxaliplatin. In patients (pts) with stage III colon cancer and deficient mismatch repair (dMMR), the benefit of an immune checkpoint inhibitor combined with adjuvant chemotherapy is unknown. The phase III ATOMIC trial (NCT02912559) was conducted to determine whether atezolizumab (atezo), an anti-PD-L1 antibody, can improve pt outcomes when added to adjuvant 5-fluorouracil, leucovorin plus oxaliplatin (mFOLFOX6) in pts with stage III dMMR tumors. Methods: We conducted an NCI-sponsored, multicenter and randomized phase III trial in pts with surgically resected stage III dMMR colon adenocarcinoma (any T, N 1,2 M 0 ). Pts, age > 12 years (yr), were accrued at NCTN sites and the German AIO. Tumor dMMR was determined by local immunohistochemistry and centrally verified. Pts were randomized 1:1 to mFOLFOX6 plus atezo (840 mg IV q2 weeks) for 12 cycles (6 months)[mo] followed by atezo monotherapy for 13 cycles (12 mo total) versus mFOLFOX6 alone for 12 cycles. Randomization stratification factors were N-stage (N 1 /N 1c vs N 2 ), T-stage (T 1 -T 3 vs T 4 ) and site (proximal vs distal). The primary endpoint was disease-free survival (DFS); secondary endpoints were overall survival and adverse event (AE) profile (CTCAE, PRO-CTCAE). Primary efficacy analysis was done in the intent-to-treat population; DFS was compared by arm (stratified log-rank test). Hazard ratio (HR) and 95% confidence interval (CI) were calculated using a stratified Cox model; 3-yr DFS was determined by Kaplan-Meier method. Among 700 pts., 165 DFS events with two interim analyses (50% , 75% of events) yielded 90% power to detect HR 0.6 (3-yr DFS 75% vs. 84.2%) assuming exponential survival and 1-sided alpha (0.025). Results: From 9/2017 to 1/2023, 712 pts were randomized (1 pediatric) to either atezo plus mFOLFOX6 (n= 355; atezo arm) or mFOLFOX6 (n= 357). Median pt age was 64 yr. 55.1% were female. Among tumors, 83.8% were proximal, 46.1% were clinical low risk (T 1-3 N 1 ) and 53.9% high risk (T 4 and/or N 2 ). At the second interim analysis, median pt follow-up was 37.2 mo (interquartile range, 24.2 to 55.5) and 124 DFS events were observed. Three-year DFS was 86.4 % (95% CI, 81.8 to 89.9) in the atezo arm and 76.6 % (95% CI, 71.3 to 81.0) in the mFOLFOX6 arm (HR, 0.50; 95% CI, 0.35 to 0.72). Stratified log-rank p-value was <0.0001, crossing the pre-specified efficacy boundary of 0.009. Efficacy for the atezo arm was consistent across subgroups, including pts >70 yr and low- and high-risk groups. Treatment-related > grade 3 AEs occurred in 71.7 % of pts in the atezo arm vs 62.1 % in the mFOLFOX6 arm. Conclusion: The addition of atezolizumab to mFOLFOX6 significantly improved DFS and should be considered the new adjuvant standard of care for patients with dMMR stage III colon cancer. Support: U10CA180821, U10CA180882, U24CA196171; Genentech, a member of the Roche group; https://acknowledgments.alliancefound.org. Clinical trial information: NCT02912559 .

3569 Background: Immune checkpoint inhibitors (ICIs) are a standard treatment for gastrointestinal (GI) cancers with mismatch repair deficiency (dMMR) or microsatellite instability (MSI). However, around 50% of patients develop resistance to ICIs, during treatment or after discontinuation. In these patients, the efficacy of reuse an ICI in patients progressing during a previous ICI (rechallenge), or in patients who progressed after discontinuation (reintroduction), remains unknown. The INFLATE study evaluates the efficacy of ICI reuse in patients who progressed on or after discontinuing ICI. Methods: This is a multicenter international retrospective study from the IMMUNODIG cohort, including patients from 34 centers in France, the United States, Italy, Belgium and Spain. All patients received ICI for dMMR/MSI GI cancer. We analyzed patients who had progressed following initial ICI (ICI-1) and subsequently received a rechallenge or a reintroduction with ICI (ICI-2), either monotherapy (mono-ICI) or biotherapy (bi-ICI). Results: A total of 77 patients were included, receiving bi-ICI (N = 34) or mono-ICI (N = 43) during ICI-2. The majority (76%) had a metastatic colorectal cancer. The reason for discontinuing ICI-1 was disease progression in 53% of cases, end of treatment in 15%, toxicity in 5% and other reasons in 26%. Patients who discontinued ICI-1 due to progression received MONO-ICI-2 in 29% of cases and BI-ICI-2 in 71%, whereas those who stopped for other reasons received MONO-ICI-2 in 86% and BI-ICI-2 in 14% of cases. Efficacy results are shown in Table 1. The ORR and DCR were 26% and 79% with MONO-ICI-2, and 16% and 71% with BI-ICI-2. Among patients who discontinued ICI-1 due to progression, BI-ICI-2 (N = 29) achieved an ORR and DCR of 8% and 65%, with a median PFS of 5.5 months (95%CI 4.07-10.4). These outcomes were 17%, 67%, and 3.7 months (95%CI 2.37-NA), respectively, with MONO-ICI-2 (N = 12). For patients who discontinued ICI-1 for reasons other than progression, BI-ICI-2 (N = 5) achieved an ORR and DCR of 60% and 100%, with a median PFS not reached (95%CI 11.5-NR), while MONO-ICI-2 (N = 31) achieved 29%, 82%, and 14.2 months (95%CI 11.2–NR), respectively. Conclusions: This is the first multicenter real-world study evaluating ICI reuse in dMMR/MSI GI cancers. In patients who discontinued ICI for reasons other than progression, reintroduction of ICI therapy upon progression achieves tumor control in 85% of cases. In cases of progression on mono-ICI, Bi-ICI re-challenge achieve tumor control in two-thirds of cases and might be to consider in some patients. ORR % DCR % PFS median (95%CI) OS median (95%CI) Overall N=77 21 76 7.65 months(5.5–11.5) 27.6 months(22.2–55.7) Discontinued ICI-1 due to Progression N=41 11 68 5.03 months(3.7–8.05) 26.7 months(21.1–NR) Discontinued ICI-1 for Other Reasons N=36 33 85 14.22 months(11.24–NR) 27.6 months(22.2–NR)

3504 Background: Among patients with resected node-positive colon cancer, nearly 30% will relapse despite standard adjuvant chemotherapy. Analysis of molecular residual disease (MRD) using circulating tumor DNA (ctDNA) may enable risk stratification for tumor recurrence and inform adjuvant treatment decisions. Methods: Postsurgical ctDNA was analyzed in patients with stage III colon carcinoma who participated in a phase 3 trial of adjuvant FOLFOX alone or combined with cetuximab (n = 3084) [NCCTG N0147]. We utilized a tissue-free epigenomic assay for ctDNA detection (Guardant Reveal) with sampling prior to start of adjuvant therapy. Among ctDNA positives, epigenomic tumor fraction (TF) was estimated and ctDNA genotyping was done with Guardant360 (panel of 739 genes). Median follow-up was 6.1 years (yr). Study endpoints included time-to-recurrence (TTR), disease-free survival (DFS) and overall survival (OS) analyzed by Kaplan–Meier method. Multivariable Cox proportional hazards models were used to assess prognostic utility of ctDNA status adjusting for confounders. Interaction between ctDNA and clinicopathological features were assessed. Results: Among 2260 patients with evaluable ctDNA data, 461 (20.4%) were ctDNA positive.Tumors were significantly associated with higher T, N stage, BRAF V600E , high grade, obstruction/perforation, and worse performance status. Positive vs negative ctDNA was significantly associated with shorter TTR (hazard ratio [HR] 4.33, 95% confidence interval [CI] 3.65-5.13, P < 0.0001), poorer DFS (HR 3.74, CI 3.18-4.39, P < 0.0001] and OS (HR 3.17, CI 2.63-3.83, P < 0.0001), adjusting for covariates and tissue MMR, KRAS and BRAF V600E . ctDNA positive vs negative patients had 3y DFS of 36.4% (95%CI 32.2-41.2%) vs 82.5% (95% CI 80.0-84.4%), respectively. Adverse prognosis was consistent across subgroups (all P < 0.05), with stronger detrimental effects for positive ctDNA in N1 (vs. N2), T1/2 (vs T3 or 4), and mismatch repair deficient tumors [interaction P = 0.0002 to 0.041). Among patients with positive ctDNA, TF in those who recurred/died within 3 yr was double of those who remained recurrence-free ( P = 0.0001). High vs. low ctDNA TF ( > vs ≤ median) further stratified TTR (HR 1.48, CI 1.17-1.88, P = 0.0011), DFS (HR 1.52, CI 1.21-1.92, P = 0.0004) and OS (HR 1.58, CI 1.21-2.07, P = 0.0009), adjusting for confounders. ctDNA positive cases, . Analyses of ctDNA detection by site of recurrence, ctDNA clearance, and genomic variant detection are ongoing. Conclusions: In the largest study evaluating tissue-free epigenomic-based MRD detection, we demonstrate a robust prognostic utility of postsurgical ctDNA. Tumor fraction provided further patient stratification and analysis is ongoing to identify a subgroup based on TF that may be unlikely to clear ctDNA despite adjuvant chemotherapy.

e15628 Background: Total neoadjuvant therapy (TNT) with cytotoxic chemotherapy and chemoradiation followed by surgical resection is a standard treatment for locally advanced rectal cancer (LARC). In a limited number of LARC patients with deficient mismatch repair (dMMR), treatment with an immune checkpoint inhibitor (ICI) resulted in all patients achieving complete clinical response (cCR) (median follow-up of 29 months) such that radiotherapy and surgery were avoided. These and other data have resulted in the routine use of neoadjuvant ICI for dMMR LARC. Here we report a multicenter series of patients with dMMR LARC treated with neoadjuvant ICI. Methods: Adult patients (N = 21) with dMMR LARC who received neoadjuvant ICI at Mayo Clinic, MD Anderson Cancer Center, Duke University or University of Pittsburgh were identified from the electronic medical record (EMR) and relevant data were abstracted. All patients were staged with pelvic MRI; clinical response was evaluated by repeat pelvic MRI and sigmoidoscopy. Data were summarized using descriptive statistics. This study was approved by the Mayo Clinic IRB. Results: Patients with clinical stage I (n = 2), II (n = 3) or III (n = 15) dMMR LARC were treated with neoadjuvant PD-1 monotherapy (n = 20) or combination (n = 1, ipilimumab+ nivolumab). Median age was 46 years (IQR 39, 65), 15 (71%) were male, 10 (48%) had distal tumors, and 14 had Lynch Syndrome. Median duration of ICI treatment was 7 mos (IQR 6, 12); one patient remains on dostarlimab at data cut-off. Clinical complete response (cCR) was achieved in 11 (52%) patients, partial response (PR) in 7 (33%), stable disease (SD) in 1 (5%), and progressive disease (PD) in 2 (10%) patients. Median time from first dose of ICI to best clinical response was 6 mos (IQR 5, 8) months. All partial or non-responders received salvage TNT with (n = 6) or without (n = 3) surgery. At a median follow up of 19 mos (range 6 to 56), tumor regrowth occurred in 1 patient who was subsequently treated with TNT. Any grade immune-related adverse event occurred in 6 (29%) patients; one grade 3 hepatitis occurred in the patient on Ipi+Nivo (Table1). Conclusions: Neoadjuvant ICI treatment of dMMR LARC resulted in half of treated patients achieving cCR. Careful monitoring is necessary to identify non responders who will require salvage TNT. Patient characteristics and clinical response to ICI. ID Sex Age Stage ICI Response to ICI ID Sex Age Stage ICI Response to ICI 1 M 62 T3N+ Pembro cPR 12 M 37 T3N+ Pembro cCR 2 M 60 T2N0 Pembro cCR 13 M 41 T3N0 Pembro cCR 3 M 80 T2N0 Pembro cPD 14 M 46 T4N+ Pembro cPR 4 M 36 T3N+ Ipi+Nivo cCR 15 M 68 T3N+ Pembro cPR 5 M 57 T3N+ Dostarli cCR 16 M 77 T3N0 Pembro cPR 6 M 39 T3N+ Dostarli cPR 17 M 40 T4Nx Pembro cCR 7 F 44 T2N+ Dostarli cPD 18 F 35 T3N+ Dostarli cCR 8 M 74 T3N+ Pembro cCR 19 M 60 T2N+ Pembro cPR 9 M 43 T3N+ Nivo cCR 20 F 65 T3N+ Pembro cCR 10 F 81 T3N0 Dorstali cPR 21 F 39 T4N+ Dorstali cSD 11 F 32 T2N+ Pembro cCR

Abstract Background: Mutant BRAF(V600E) is present in 8-12% of human colorectal cancers (CRCs) and is associated with treatment resistance and poor survival. Targeted therapy with inhibitors of BRAF and EGFR is associated with only an incremental survival benefit. Methods: We studied the novel CDK2/CDK9 inhibitor, fadraciclib, alone and in combination with the BRAF inhibitor encorafenib on isogenic CRC cells with two, one, or no BRAFV600E alleles (RKO+/+, A19+/-, T29-/-). Analysis of apoptosis, colony formation, and the cell cycle was performed. The effect of drug treatment on growth of tumor xenografts was determined in a zebrafish model. Results: Fadraciclib induced apoptosis and inhibited colony formation in a gene dose-dependent manner. Its growth inhibitory effect when combined with encorafenib was synergistic. Mechanistically, fadraciclib suppressed the expression of pro-survival MCL-1 and decreased the phosphorylation of RNA polymerase II indicating suppression of RNA transcription. Fadraciclib also impaired Rb phosphorylation and promoted a DNA damage response indicated by the cleavage of PARP, pH2AX (ser139), and activation of p53. Drug treatment was associated with a greater suppression of Rb phosphorylation, cell cycle inhibition, and apoptosis in RKO+/+ versus A19+/- and T29-/- cells. In a zebrafish tumor xenograft model, fadraciclib plus encorafenib significantly reduced the tumor xenograft size concurrent with increased caspase-3 activation. Conclusion: Fadraciclib plus encorafenib potently induced apoptosis and synergistically suppressed tumor cell growth in a gene dose-dependent manner in BRAFV600E CRC cells, and exerted anti-tumor effects in a tumor xenograft model. These data suggest a novel therapeutic strategy in this CRC subtype. Citation Format: Md Mohiuddin, Vanda Póvoa, Rita Fior, Frank A. Sinicrope. Novel CDK2/CDK9 inhibitor fadraciclib targets cell survival and DNA damage pathways and synergizes with encorafenib in human colorectal cancer cells with BRAF(V600E). [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 1682.

Abstract Purpose: KRAS G12C inhibitors can treat KRASG12C-mutant advanced colorectal cancers and pancreatic ductal adenocarcinomas (PDAC), but alterations in Kirsten rat sarcoma (KRAS), EGFR, BRAF, MAP2K1, and other genes bypass KRAS inhibition and reduce therapy efficacy. Our study evaluates the genetic landscape of candidate primary resistance alterations relevant to KRAS targeting in KRASG12C-mutant colorectal cancer and PDAC. Experimental Design: We analyzed two cohorts (national database and Mayo) of patients with advanced colorectal cancer or PDAC tested with next-generation sequencing of ctDNA via Guardant360. Cohorts were divided into three groups: KRASG12C alone (KRASG12C without a resistance gene), KRASG12C with resistance (KRASG12C and ≥1 candidate resistance gene), and KRAS not detected. Candidate resistance mutations were inferred from the reported literature. Results: Among the national (13,603 colorectal cancer and 5,016 PDAC cases) and Mayo (741 colorectal cancer and 422 PDAC cases) cohorts, resistance alterations were identified in a considerable number of KRASG12C cases (46.5% of national colorectal cancer, 16.4% of national PDAC, 53.8% of Mayo colorectal cancer, and 36.4% of Mayo PDAC). The presence of resistance alterations was associated with a trend toward worse overall survival in KRASG12C colorectal cancer (P = 0.05). Conclusions: Putative resistance alterations are prevalent in PDAC and colorectal cancer and may limit monotherapy efficacy. Identifying these alterations has potential implications in optimal patient selection for targeted therapies and the development of combination therapy strategies to overcome primary resistance.

Abstract Purpose: Digital proteomic profiling was performed to identify spatial context in relationship to patient response and survival after anti–PD-1 therapy in metastatic colorectal cancer. Experimental Design: Primary colorectal cancers with deficient mismatch repair from patients treated with anti–PD-1 antibodies were analyzed (N = 30) using digital spatial profiling (GeoMx nCounter). At the invasive margin, 71 proteins were profiled in 10 regions of interest/slide that were segmented into 3 compartments labeled with pan-cytokeratin (epithelia), CD45 (stromal cells), and SYTO13 (nuclei). In an independent cohort (n = 13), digital spatial profiling data and single-cell transcriptomic data were analyzed. Differential protein abundance, after Benjamini–Hochberg correction, was examined by clinical response and progression-free survival (PFS) using multivariable Cox regression. Results: Protein abundance varied significantly between epithelial and stromal compartments. Nonresponders to anti–PD-1 showed higher fibronectin and smooth muscle actin abundance in the epithelial compartment that was associated with significantly shorter PFS (adjusted HR: 6.49 and 4.52, respectively; P &lt; 0.05). In CD45+ stroma, increased expression of proteins related to T cells (CD3 and CD4), NK cells (CD56), antigen presentation (CD40), immune activation (CD27, ICOS), and apoptosis (GZMA) were found in responders (vs nonresponders) to anti–PD-1; each marker was significantly associated with longer patient PFS (0.02 &lt; adjusted HR &lt; 0.17; P &lt; 0.05). In a separate cohort, consistent results by compartment were found for fibronectin and CD56. Gene expression data revealed that fibronectin and smooth muscle actin were primarily derived from cancer-associated fibroblasts. Conclusions: Spatially resolved protein profiles within microenvironments of deficient mismatch repair colorectal cancers can influence patient response and survival after anti–PD-1, highlighting their potential clinical significance.

484 Background: The histopathological growth patterns (HGPs) of liver metastasis reflect the complicated and varied interactions between tumor cells and the host microenvironment. We studied HGPs of gastric cancer liver metastasis (GCLM) and sought to predict them using radiomic and clinical variables. Methods: Hepatectomy was performed in 62 patients with GCLM at our university hospital between 2011 and 2021. HGPs were evaluated according to international consensus guidelines, and were classified as either desmoplastic HGP (dHGP) or non-dHGP. CD4, CD8, VEGF and CD31 were analyzed by immunohistochemistry. Eight hundred and fifty-one radiomics features were extracted from CT portal venous phase images and the optimal radiomics signature was determined by univariate analysis and LASSO regression. Multivariable regression analyses and ROC curves were used to assess the predictive performance of HGPs with a radiomic and clinical combined model. Results: Among 62 patients, 14 were categorized as dHGP, and 48 as non-dHGP. Non-dHGP vs. d-HGP was associated with higher tumor N stage and CEA level, and with decreased CD8 + T cells, and lower VEGF and CD31 expression. HGP status was independently associated with patient cancer-specific survival (CSS) (P＜0.05) multivariately.Analysis of percent relative contribution revealed that HGP ranked first for CSS (35.9%), ahead of T (28.9%) and N (28.9%) stage. A model with combined radiomic and clinical features demonstrated robust performance with area under the curve (AUC) values of 0.993 and 0.933 in the training and validation sets, respectively. Conclusions: Non-invasive determination of radiomic and clinical features was shown to predict HGPs types. Compared to dHGP, GCLM with non-dHGP exhibited significantly lower immune cell infiltration and decreased angiogenesis. Among features examined in GCLM, HGP was the most robust for prognostication. Univariate and multivariate analysis of prognostic factors for cancer-specific survival. Characteristics Univariate analysis Multivariate rergression analysis HR (95%CI) P HR (95%CI) P Gender ( Women vs Men ) 1.13(0.61-2.09) 0.71 Age ( >60years vs ≤60years) 1.08(0.60-1.70) 0.98 HGP ( non-desmoplastic HGP vs desmoplastic HGP ) 3.63(1.78-7.41) 0.00 3.48(1.37-8.88) 0.01 T stage ( T3-T4 vs T1-T2 ) 2.03(1.08-3.80) 0.03 2.29(1.15-4.58) 0.02 N stage ( N1-N3 vs N0 ) 3.46(1.66-7.23) 0.00 2.54(1.17-5.56) 0.02 CEA ( >5ng/ml vs ≤5ng/ml ) 2.37(1.25-4.48) 0.01 1.56(0.70-3.48) 0.27