Association of COVID-19 Lockdown With the Tumor Burden in Patients With Newly Diagnosed Metastatic Colorectal Cancer

This cohort study assesses the association of local and national COVID-19 restrictions with tumor burden, as measured by total plasma circulating tumor DNA concentration, among patients newly diagnosed with metastatic colorectal cancer.


1.
CirDNA analysis All blood samples were collected in 10-milliter (mL) Streck tubes in all 18 clinical centers that sent by express mail the day or the day after blood draw following strict procedural guidelines. The IRCM U1094/ INSERM laboratory received blood samples within 5 days post-blood draw and centralized all following procedures towards cirDNA analysis under stringent pre-analytical guidelines reported by the laboratory 24 . The blood was then centrifuged at 1,200 g at 4°C for 10 minutes. The supernatants were isolated in sterile 1.5 mL Eppendorf tubes and centrifuged at 16,000 g at 4°C for 10 minutes. Afterwards, the plasma was either immediately used for DNA extraction or stored at -20°C. CirDNA was extracted from 1 mL of plasma using the QIAmp DNA Mini Blood kit (Qiagen) according to the ''Blood and body fluid protocol.'' DNA extracts were kept at -20°C until used 25 .
The PANIRINOX clinical study implies the use of the IntPlex methodology to select patients showing WT mutational status for KRAS, NRAS and BRAF genes. IntPlex was established by our team and is based on an allele-specific blocker Q-PCR-based method specific for cirDNA analysis to enable the detection of point mutation and to determine mutant allele concentration 20 . This method combined the use of (1) allele-specific Q-PCR with blocking 3′-phosphate-modified oligonucleotide; (2) low Tm primers with mutation in 3′; (3) an integrated primer design; (4) routine internal positive and negative controls, and (5) optimal analytical procedures. This method enables the determination of the presence of a mutation, as well as the total concentration of cirDNA, the concentration of mutant cirDNA, the mutant allele frequency, and an index of DNA integrity. Note, the total cirDNA concentration value is determined by targeting a KRAS WT sequence, and is internally controlled by targeting a BRAF WT sequence for each sample 25 .
CirDNA quantification methodology and the data description were carried out according to the MIQE guidelines 20 . Q-PCR amplifications were carried out at least in duplicate in a 25-μl reaction volume on CFX96 thermocycler (Bio-Rad) using the Bio-Rad CFX manager. Each PCR mixture was composed of 12.5 μl of PCR mix (Bio-Rad Supermix SYBR Green), 2.5 μl of each amplification primer (0.3 pmol/μl), 2.5 μl of PCR-analyzed water or 2.5 μl of oligoblocker, and 5 μl of DNA extract. Thermal cycling consisted of three repeated steps: a 3-minute hot-start polymerase activation-denaturation step at 95°C followed by 40 repeated cycles at 95°C for 10 seconds and then at 60°C for 30 seconds. Melting curves were obtained by increasing the temperature from 55 to 90°C with a plate reading every 0.2°C. In the PANIRINOX study, 28 different mutations on the KRAS, BRAF and NRAS genes actionable in mCRC management care are tested. The IntPlex assay is clinically validated 20 , 21 , and shows the highest sensitivity and specificity so far described 8 . Poisson law experiments demonstrated that this method could accurately detect down to one molecule per PCR reaction mixture 21 , with detection sensitivity as high as 1/100,000.
Intra-and inter-reproducibility experiments combining pre-analytic and analytic procedures demonstrated that the coefficient of variation for cirDNA concentration measurement is 24% 25 . IntPlex has been validated in two clinical studies 20 , 21 is currently involved in 9 others, and has already enabled the testing of more than 2,000 individuals and 4,500 plama samples.

Ad hoc retrospective study
The increase seen in cirDNA values post-lockdown is striking, and points to levels of tumor burden at diagnosis which have been shown to affect patient survival 38,39 . To estimate the potential impact on survival of diagnostic delays arising from the lockdown, we retrospectively analyzed data accumulated during our last two clinical studies 21,22 from which we identified "all comers" newly diagnosed patients with mCRC , and in which a rigorously identical methodology was used to assess cirDNA (N=135) before patients began first line chemotherapy. The median cirDNA concentration (24.4 ng/mL, range [2.3-1406], Suppl. Table 2) of these patients is similar to that observed in the prelockdown cohort studied from the PANIRINOX trial. As observed in Figure 3A, (Figure 3C, Suppl. Table 2). The comparative study of the lockdown impact ( Figure 1) revealed 23 out of 40 patients (58%) showing a cirDNA amount over 100 ng/mL post-lockdown, and only 6 out of 40 patients (15%) above this concentration in pre-lockdown period. This could mean that 17 (23 minus 6) or 43% (58% minus 15%) supplementary patients diagnosed post-lockdown would have a median survival which is 54% (from 19.3 to 8.8 ng/mL) less than that of patients diagnosed pre-lockdown, presuming care management was equivalent, despite the difference in time period between survival cohort and study cohort. We are aware that this remains an assumption, but one which nonetheless illustrates and anticipates the lockdowns' marked deleterious impact on these patients' health.

Illustration of pandemic related anxiety by a cancer patient quote.
From a sixty year-Australian old patient, currently being treated for jaw cancer, who said: "Because I didn't think it was an emergency, I didn't make any moves to go back to see anybody. And also because it was in the middle of COVID I was afraid of getting COVID or giving COVID, and so I just put up with it for a while." 31

4.
Statistics Statistical analysis of pre-versus post-lockdown data was performed using the GraphPad Prism V6.01 software and survival analysis with STATA 16.0 software. Where appropriate, data were log transformed prior to statistical analysis. Continuous variables were compared using the Mann-Whitney test. Categorical variables were compared using the Pearson's chi-square test. Median follow-up was calculated using the reverse Kaplan-Meier method. Overall survival (OS) was estimated using the Kaplan-Meier method, and compared using the Log-rank test. OS was defined as the time between the date of first metastatic diagnosis and the date of death from any cause. Hazard ratios (HR) are given with their 95% confidence interval (95% CI). Correlation analysis were performed using the spearman test. A probability of less than 0.05 was considered to be statistically significant; *p<0.05, **p<0.01; ***p< 0.001; ****p< 0.