The COVID-AGICT study: COVID-19 and advanced gastro-intestinal cancer surgical treatment. A multicentric Italian study on the SARS-CoV-2 pandemic impact on gastro-intestinal cancers surgical treatment during the 2020. Analysis of perioperative and short-term oncological outcomes

On March 9, 2020, Italy imposed a national lockdown to curb COVID-19, reallocating health resources toward infected patients. These measures disrupted cancer screening, diagnostics, elective oncologic treatments, and emergency surgery, increasing preoperative delays and potentially avoidable cancer-related deaths. Italy, among the earliest and hardest-hit Western countries, had limited data on the disruption of surgical services for oncologic patients. This multicentric national study aimed to investigate changes in outcomes of patients undergoing surgery for colorectal, gastroesophageal, and pancreatic cancers, comparing the year before the pandemic (2019) with the first pandemic year (2020). The primary research question was whether the pandemic increased the proportion of advanced-stage resections in 2020 compared with 2019, with secondary aims to assess perioperative and short-term oncologic outcomes.

The authors reference international and Italian reports indicating widespread disruptions in cancer care during COVID-19, including increased diagnostic delays, reduced elective surgery, and concerns about long-term oncologic outcomes. Several studies reported unchanged short-term pathologic stages during the pandemic, though some (e.g., multicenter studies from northern Italy and England) found more advanced clinical features at presentation and shifts toward neoadjuvant therapies (e.g., increased radiotherapy for rectal cancer). Guidance documents debated the safety of minimally invasive surgery early in the pandemic, but later evidence found no clear increased risk of viral transmission via surgical smoke compared with open surgery. Large international cohorts (COVIDSurg) documented increased preoperative delays and higher nonoperation rates during lockdowns.

Design: National multicentric retrospective cohort study across 62 Italian surgical divisions. Period: January 2019 to December 2020. Registration and Ethics: Approved by coordinating center ethics committee (ref. 18886); registered at ClinicalTrials.gov (NCT04686747). Consent waived due to retrospective, anonymized data. Population: Adult patients undergoing surgery for colorectal, gastroesophageal, or pancreatic cancers (localized, locally advanced, or metastatic), with curative or palliative intent. Exclusions: age <18, multiple tumors, not surgically treated, or hospitals without complete data for both years. Data collected: Demographics (BMI, age, sex, ASA, Charlson Comorbidity Index [CCI]); preoperative outcomes (urgent diagnosis due to symptomatic cancers; use of neoadjuvant treatments [NCTs]; interval from diagnosis to operation in days, excluding NCT patients for this metric); perioperative outcomes (urgent colorectal cancer resection, unresectability, minimally invasive surgery [MIS] rate and conversion, length of stay [LOS], major postoperative complications per Clavien-Dindo III–V, 30-day readmission, 30-day mortality); early oncologic outcomes for curative-intent cases (adjuvant therapy, number of lymph nodes retrieved, proportion of positive nodes, proportion of node-positive patients, margin status R1/R2, lymphovascular and perineural invasion). Primary endpoint: Proportion of advanced-stage surgeries in 2020 vs 2019. Staging definition used AJCC 8th edition: early stage (non-nodal, non-metastatic, R0 resection) vs advanced (nodal-positive, metastatic, and/or R1–R2 resection). Secondary endpoints: Impact of COVID-19 on perioperative and postoperative outcomes. Time-based analysis divided each year into quarters (Q1: Jan–Mar; Q2: Apr–Jun; Q3: Jul–Sep; Q4: Oct–Dec) to compare surgical activity, advanced-stage resection rate, and frequency of palliative/urgent procedures. Statistical analysis: Quantitative variables reported as mean ± SD; compared using two-tailed Student’s t-test (accounting for heteroskedasticity). Qualitative variables as counts and percentages; compared with Cochran–Mantel–Haenszel chi-square extension for multicentric data, also used for quarterly distributions. Significance threshold p ≤ 0.05. Missing data handled listwise; patients with >25% missing for a specific outcome were excluded from analyses of that outcome. Proportions of available data per outcome reported.

- Cohort: 8250 patients; 4370 (53%) in 2019 and 3880 (47%) in 2020 from 62 centers. Cancer types: 5704 colorectal (69%), 1816 pancreatic (22%), 730 gastroesophageal (9%). Over half of centers were in northern Italy. - Primary endpoint: Advanced-stage resections were 51% in 2020 vs 49% in 2019; difference not significant (P = 0.25; 90% data available). In colorectal subgroup, 46% vs 44% (P = 0.23; 93% data available). - Quarterly analysis: In Q3–Q4 2020, advanced cancers resected tended higher than same quarters 2019 (50% and 49% vs 48% and 45%; P = 0.05). Palliative and urgent procedures also increased in late 2020 vs 2019 but not significantly (P = 0.6 and P = 0.7). - Demographics: No differences in age, sex, BMI, ASA. Higher CCI score in 2020 (5.38 ± 2.08) vs 2019 (5.28 ± 2.22), P = 0.04 (91% data). - Preoperative: Urgent diagnosis more frequent in 2020 (24.2%) vs 2019 (20.3%), P < 0.001 (100% data). Neoadjuvant treatments increased (24.1% vs 19.5%), P < 0.001 (91% data); rectal cancer NCTs increased (389 vs 365 patients, P < 0.001; 87% data). Time from diagnosis to surgery increased (64.2 vs 56.8 days), P < 0.001 (90% data). - Surgical activity: Overall 5.2% reduction in 2020 (P < 0.001; 91% data). Q1 2020 −8% (P = 0.04); Q2 2020 −29% (P < 0.001) vs 2019. - Perioperative: Urgent colorectal resections increased (9.3% vs 7.3%), P = 0.005 (100% data). Unresectable cases similar (6.4% vs 5.9%), P = 0.28 (95% data). MIS overall similar (62.1% vs 63.4%), P = 0.23 (99% data). Overall conversion rate lower in 2020 (7.2% vs 9.2%), P = 0.01 (94% data). In colorectal subgroup: MIS rate lower (75.6% vs 78.4%), P = 0.009 (99% data); conversion lower (6.9% vs 8.7%), P = 0.03 (93% data). - Postoperative: LOS shorter in 2020 (11.6 ± 11.0 days) vs 2019 (12.1 ± 11.4), P = 0.04 (99% data). Major complications (Clavien-Dindo III–V) similar (12.8% vs 13.7%), P = 0.20 (97% data). 30-day readmission: 5.0% vs 4.4%, P = 0.16 (95% data). 30-day mortality: 2.5% vs 1.9%, P = 0.09 (96% data). - Early oncologic outcomes (curative intent): Adjuvant therapy rates similar (36.2% vs 35.9%), P = 0.80 (77% data). Lymph nodes retrieved fewer in 2020 (24.6 vs 25.4), P = 0.03 (89% data). Proportion of positive nodes per surgery higher in 2020 (9% vs 7%; 2.22 ± 4.63 vs 1.96 ± 4.18), P < 0.001 (92% data). Proportion of node-positive patients similar (43% vs 41.6%), P = 0.23 (92% data). R1/R2 margin rates similar (7.4% vs 7.9%), P = 0.4 (91% data). Lymphovascular/perineural invasion similar (63.6% vs 62.1%), P = 0.25 (75% data).

The study did not find a statistically significant increase in advanced pathological stage at surgery in 2020 versus 2019 for colorectal, gastroesophageal, and pancreatic cancers, addressing the primary question that stage migration at the time of operation was not evident overall. However, quarterly analyses suggest a trend toward more advanced resections in the second half of 2020, coinciding with delayed care and recovery from initial lockdown effects. The pandemic environment was associated with higher comorbidity burden (CCI), more urgent presentations, increased use of neoadjuvant therapies (particularly for rectal cancer), longer preoperative delays, and higher rates of urgent colorectal resections. These changes likely reflect disruptions in screening and elective pathways, more symptomatic disease necessitating emergency care, and strategic use of neoadjuvant therapy as a bridge to surgery during resource constraints. Operative practice adapted: while overall MIS use remained similar, colorectal MIS decreased, possibly due to early caution regarding laparoscopy and an increase in emergencies, but conversion rates fell, suggesting stricter selection and optimization for MIS cases. LOS was shorter in 2020, potentially reflecting institutional drive to minimize hospital exposure and patient preference to expedite discharge. Early oncologic metrics were largely stable, though a higher proportion of positive nodes per surgery in 2020 may represent an early signal of more biologically advanced disease at the time of resection influenced by delays and urgent pathways. These findings align with international reports documenting reduced surgical activity and increased preoperative delays during lockdowns. While short-term pathological staging did not worsen overall, the shifts in perioperative metrics and nodal positivity raise concerns for potential longer-term oncologic consequences as the pandemic progressed.

In this multicentric Italian study, the COVID-19 pandemic did not significantly change overall pathological staging at surgery for colorectal, gastroesophageal, and pancreatic cancers in 2020 compared with 2019. Nonetheless, 2020 saw higher comorbidity burden, increased urgent diagnoses and urgent colorectal resections, greater use of neoadjuvant treatments, longer delays from diagnosis to surgery, shorter hospital stays, and a higher proportion of positive lymph nodes, with a trend toward more advanced resections during the second half of 2020. These results highlight the need to preserve dedicated diagnostic and treatment pathways for gastrointestinal cancers during public health crises to maintain care standards and mitigate downstream oncologic impacts.

- Retrospective design with inherent potential for selection and information bias. - Non-operated patients in 2020 were not captured, limiting assessment of full epidemiologic impact and potentially underestimating stage migration. - Some variables lacked detail, and missing data handling (listwise exclusion) may affect certain estimates. - Heterogeneity across five gastrointestinal cancer types may confound perioperative and postoperative outcomes due to differing biology, treatment indications (e.g., neoadjuvant therapy), and complication profiles. - The two-year window (2019–2020) may be too short to detect definitive differences in pathological staging; inclusion of 2021 could yield additional insights. - The cohort was dominated by colorectal cancers (69%), which may influence overall outcomes and generalizability. - Despite inclusion of 62 centers and 8250 patients, findings may not fully represent the nationwide impact on GI surgical oncology.