Molecular Profiling and Prognostic Value of KRAS Gene Mutations in Tumor Samples and Liquid Biopsies from Patients with Pancreatic Ductal Adenocarcinoma

Morão B¹, Fernandes AR²

Introduction:

assess KRAS mutations and its prognostic value in PDAC patients.

Methods: Prospective multicentric and overall survival (OS).

Results: Of 55 months, p=0.02) and lower OS (1 vs 8 months, p<0.01).

Conclusions:

Keywords

tion system, ARMS/HRMA, prognosis

Key Summary

• of PDAC patients • • • - - - -

Introduction

Current trends suggest that pancreatic cancer (PC) will become the second leading cause of cancer-related deaths and the most lethal digestive cancer by 2030 [1,2], with pancreatic ductal adenocarcinoma (PDAC) accounting for 85% of cases. The 5-year survival rate is only 9% if all stages combined [3]. This discouraging scenario is mainly explained by a late diagnosis and the lack of an effective therapy. Surgical resection followed by adjuvant chemotherapy (ChT), the only potentially curative therapy, is only possible in a minority of patients. Novel biomarkers to be used as diagnostic, predictive tools of response and targets to tailored therapy, may contribute to improve the prognosis [4].

KRAS mutation occurs early in pancreatic carcinogenesis, and is present in at least 80% of PDAC [5,6]. Over 90% of activating mutations are found at codon 12 of the oncoprotein (p.G12D, p.G12V, p.G12R, pG12C or p.G12A) corresponding to single-nucleotide mutations at the cDNA sequence (c.35G>A, c.35G>T, c.34G>C, c.34G>T or c.35G>C, respectively) [7,8]. Although there are mixed findings in the literature, several studies have already demonstrated the prognostic value of KRAS mutations in tumor samples and liquid biopsies in patients with PDAC [9]. Tumor KRAS G12D mutation has been shown to be an independent predictor of worse overall survival in patients with PDAC, while KRAS G12V mutation was associated with better tumor regression following chemotherapy and radiotherapy in borderline and locally advanced PDAC [10-12]. However, there are conflicting findings and KRAS G12V-mutant tumors have also been proposed has having worse outcomes than other subtypes [13,14]. Additionally, the presence and levels of circulating tumor DNA (ctDNA) harboring KRAS mutations have been shown to be a predictor of worse overall survival (OS) in patients with PDAC and changes in its concentration during ChT were better to predict response to treatment than changes in CA 19.9 [15-18]. In one of these studies, levels of ctDNA correlated with disease status, with higher levels found in PDAC patients than in healthy controls; furthermore, ctDNA levels lowered after resection surgery, suggesting that ctDNA levels may reflect disease burden [18]. Most recently, treatment with a KRAS G12C selective inhibitor in patients with locally advanced or metastatic PDAC resulted in clinically meaningful responses (objective response 21%, disease control 84%, median duration of response 6 months) [19], adding to the importance of assessing KRAS mutational status in PDAC.

Despite recent advances, the detection ctDNA is still demanding due to its low abundance in peripheral blood, and high-sensitivity molecular techniques are needed [20]. The lack of standard methods for detection of ctDNA readily available in clinical practice and at a low-cost, have prevented wider use of ctDNA technology [21].

The purpose of this study was: 1) To further validate a single-step methodology based on Amplification-Refractory Mutation System (ARMS) coupled with High-Resolution Melting Analysis (HRMA) to assess codon 12 KRAS mutations in a cohort of PDAC patients with stored tumor and plasma samples 2) To assess its prognostic value regarding disease progression and all-cause mortality.

Materials and Methods

Study design and inclusion criteria

We conducted a longitudinal, multicentric, prospective cohort study including patients with newly diagnosed and histologically proven PDAC treated at the referral center Hospital Beatriz Ângelo (Loures, Portugal) and Hospital da Luz (Lisbon, Portugal), between October 2017 and May 2021. According to tumor staging and performance status (PS) patients were either referred for direct surgery or for ChT. In the latter, an endoscopic ultrasound-guided fine-needle biopsy (EUS-FNB) was performed to confirm the histological type of the tumor.

We included adult patients (18 years old or older) with newly diagnosed PDAC who were treated in both hospitals, with available tumor and blood samples collected before therapy and with the ability of giving written informed consent. Patients previously treated with surgery, chemotherapy or radiotherapy; those referred to neoadjuvant or palliative chemotherapy but with a biopsy performed at another institution; those unable to undergo surgery or EUS-FNA for histologically proven PDAC; and those who underwent direct surgery with a pancreatic tumor other than PDAC were excluded from this study. The study protocol was approved by both Hospitals ethical committees. Informed written consent was obtained from all the patients.

During the study period, 256 patients with PDAC were treated at both institutions (Figure 1). We were able to collect tumor and plasma samples at diagnosis from 95 patients that were stored for later analysis. Of these, 55 PDAC patients had available fresh frozen tumor and plasma samples and were included in the present study. Tumor samples were collected during surgery in 28 patients (patients referred for direct surgery) and by EUS-FNB in 27 patients (patients referred for neoadjuvant ChT). We could extract tumor DNA for mutational analysis by SS in all 55 cases, but we did not have enough tumor DNA sample left for further ARMS/ HRMA analysis in 11 patients due to cell paucity (6 samples had been collected by EUS-FNB and 5 samples during surgery).

SS of the tumor sample was performed in all 55 patients. For comparison and validation of ARMS/HRMA technology, the 2 most frequent KRAS mutations in codon 12 (G12D and G12V) were analyzed. As 11 patients were excluded from analysis because there was no tumor DNA sample left, and 6 out of 55 samples had

a different G12R or G12C mutations, a population of 38 patients was used to compare the performance of SS and ARMS/HRMA technology.

Tumor sample collection

A tumor sample was collected at diagnosis before any treatment, in one of the following settings: 1) during resection surgery in patients who underwent direct surgery, or 2) during EUS-FNB in patients undergoing neoadjuvant or palliative ChT as first-line treatment. During the preparation of the surgical specimen, a 5 mm of the tumor mass was collected into a sterile 1-ml Eppendorf kept at 4ºC, immediately frozen in liquid nitrogen and stored at -80ºC for later molecular analysis. EUS was performed under intravenous propofol anesthesia using a curved linear array echoendoscope by Olympus connected to Aloka ultrasound device. EUS-FNB was performed using a 22-gauge needle. A minimum of two needle passes were performed for diagnosis. Sample adequacy was either evaluated by Rapid On-site Evaluation by a pathologist at the endoscopy room or by Macroscopic On-Site evaluation. When the sample obtained was considered to be inadequate for diagnosis, an extra needle pass was performed to collect material for the study, never exceeding a total of five passes in order avoid adverse events. Core-biopsy samples were transferred into Cytolyt media and were used for histological diagnosis according to usual clinical practice.

Blood sample collection

Blood samples were collected at diagnosis before any treatment. A small amount of blood (8mL) was collected into 2 EDTA blood tubes and immediately kept at 4ºC; they were transported to the laboratory at 4ºC in a maximum period of 4 hours (ideally 2 hours). Blood components were then separated by centrifugation at 1000G at 4ºC and placed in cryotubes: 4 containing plasma and 1 containing red blood cells. Blood components in cryotubes were fresh frozen in liquid nitrogen and stored and kept at -80ºC for later analysis.

KRAS mutation analysis

Materials and reagents: For DNA Isolation the High Pure PCR Template Preparation Kit (Roche, Basel, Switzerland) was used. MgCl , dNTPs, NZYTaq II DNA Polymerase and NZYSpeedy qPCR Green Master Mix (2x) were purchase from NzyTech, Lisbon, Portugal. DMSO was purchase from Sigma-Aldrich, St. Louis, Missouri, USA. The DNA size marker GeneRulerTM DNA Ladder Mix was purchased from Fermentas (Burlington, Canada).

® GelRed Nucleic Acid Gel Stain was purchased to Biotium (California, USA). All the primers were purchased from STAB VIDA, Lda (Caparica, Setúbal, Portugal).

DNA Extraction: DNA extraction from PDAC patients’ tumor and plasma samples was performed using the High Pure PCR Template Preparation Kit (Roche, Basel, Switzerland) according to the manufacturer’s instructions (from solid tissue and whole blood/plasma). AASSElution volume was altered to 50 μL for tumors and 30 μL for plasmas. DNA purity and yield were quantified using a Nanodrop spectrophotometer (Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA).

PCR amplification to assess KRAS exon 2 mutational status

with SS: S50 ng of tumor/ plasma DNAs were used for PCR amplification using 0.12 μM of each primer (KRAS forward: 5´-GGTG- GAGTATTTGATAGTGTA-3´; KRAS reverse: 5´- TGGACCCT- GACATACTCCCAAG -3´), 2 mM of MgCl (NzyTech, Lisbon, Portugal), 0.8 mM of dNTPs Mix (NzyTech, Lisbon, Portugal) and 0.15 U of NZYTaq II DNA Polymerase (NzyTech, Lisbon, Portugal) for a final volume of 20 μL of reaction. Reactions were performed on a MyCycler Thermocycler (Bio-Rad, Hercules, California, USA) using a denaturing step of 95°C for 5 min, followed by 30 cycles of 95°C for 30 s, 61°C for 30 s for tumor and 53ºC for 30 s for plasma samples, and 72°C for 20 s. For each plasma samples four independent amplification reactions were performed, concentrated in a SpeedVac (Thermo Fisher Scientific, Waltham, Massachusetts, USA) and resuspended in 15 μL of DEPC-treated water. PCR products were then direct sequenced by STAB VIDA (Caparica, Setubal, Portugal), and the chromatograms analyzed using FinchTV software (Geospiza, Inc) for sequence characterization and identification of possible mutations in exon 2.

ARMS-HRMA for detection of KRAS G12V and G12D mu-

tations: Tumor and plasma DNA samples were analyzed for the G12D and G12V KRAS point mutations through ARMS/HRMA (amplicon size 96 bp). The reaction mixture consisted of 4 ng of template DNAs, 0.3 μM of allele specific forward primers and a common reverse primer (KRAS Forward G12V 5´-CTTGTGG- TAGTTGGAGCTTT-3´; KRAS Forward G12D 5´-CTTGTGG- TAGTTGGAGCTTA-3´; KRAS Reverse 5´-CTCTATTGTTG- GATCATATTCG-3´), 2% (v/v) of DMSO (Sigma-Aldrich, St. Louis, Missouri, USA), 1x of NZYSpeedy qPCR Green Master Mix (2x) (NzyTech, Lisbon, Portugal) in a final reaction volume of 10 μL.

ARMS-HRMA reactions were performed on a Corbett Rotor-Gene 6000 (Qiagen, Hilden, Germany) thermocycler. Briefly, an initial denaturation at 95°C for 3 min, 10 cycles of 95°C for 30 s, 52°C for 15 s for G12V mutation and 54°C for 15 s for G12D mutation, and 72°C for 10 s; followed by 25 cycles of 95°C for 30 s, 60°C for 45 s, and 72°C for 10 s. After ARMS reaction, amplification products were assessed by HRMA with an initial increase of temperature from 45°C to 95°C, rising at 0.2°C per step/wait 5 sec each step. The resulting derivative plot was generated using Rotor-Gene 6000 Series Software 1.7 (Qiagen, Hilden, Germany).

Variables and endpoints

Baseline demographic and clinical data included gender, age, date of diagnosis, Eastern Cooperative Oncology Group (ECOG) performance status, serum CA 19.9 level, disease stage according to American Joint Committee on Cancer (AJCC) TNM classification and resectability status according to National Comprehensive Cancer Network (NCCN). During follow-up, the following variables were collected: occurrence and date of disease progression (either locally or metastatic) detected on computed tomography or magnetic resonance cholangiopancreatography, and date of last follow-up or death. The primary endpoint was the frequency of codon 12 KRAS mutations detected in the primary tumor and liquid biopsies of patients with PDAC by ARMS/HRMA technology and its concordance with Sanger sequencing. The secondary endpoints were the association of KRAS mutations in tumor and liquid biopsies with the occurrence of disease progression and death during

follow-up. Disease progression was defined by radiological and/ or serologic criteria. The former was considered as the appearance of new infiltrative disease on CT or MRI during follow-up, either locally or metastatic. The latter was defined as CA 19-9 elevation in association with clinical deterioration, despite no radiological evidence of disease progression.

Follow-up and treatment

After histological confirmation of PDAC, patients were routinely managed according to the standard of care and international guidelines. The KRAS status in tumor and plasma samples of each patient was only evaluated later, and as such, did not influence therapeutic decision.

Statistical analysis

Data was processed and analyzed with IBM SPSS Statistics ® ® 22 (Statistical Package for Social Sciences, IBM, Armonk, NY). Categorical variables were described through absolute and relative frequencies and continuous variables as mean and standard deviation, median, minimum, and maximum. Hypotheses about the distribution of categorical variables were tested by using the Chisquare test. Hypotheses about the distribution of continuous variables were tested by using the independent sample t-test/one-way Anova or nonparametric Mann-Whitney and Kruskal-Wallis test, depending on if normal or non-normal distribution, respectively, and considering the nature of the hypothesis. All hypotheses were tested at 5% level of significance. Survival analysis was performed between the group was performed using the log-rank test. Overall survival (OS) was defined as the length of time from the date of diagnosis that patients are still alive, taking into consideration death by any cause. Progression free survival (PFS) was defined as the length of time during and after treatment that patients lived with the disease, until the occurrence of either disease progression or death.

Results

Population baseline characteristics

Patients’ characteristics according to tumor KRAS genotype are shown in Table 1. Among the 55 patients included, 32 (58%) presented a mutation in codon 12 of KRAS protein by SS. The most frequent mutation was c.35G>A (G12D; 15 patients; 27%), followed by c.35G>T (G12V; 11 patients; 20%), c.34G>C (G12R; 4 patients; 7%) and c.34G>T (G12C; 2 patients; 4%). No other KRAS mutations were found. Baseline characteristics did not differ between patients with wild-type and mutated KRAS, neither between the different KRAS mutations, namely sex distribution, age, performance status score, TNM stage, clinical stage, and general treatment (Table 1).

Patient management according to clinical staging is summarized in Figure 2. All 7 patients with borderline resectable disease underwent neoadjuvant ChT with FOLFIRINOX regimen. Of these patients, only 3 (43%) were operated; the remaining four were kept on palliative ChT because of disease progression or because KRAS-wt – KRAS wild-type; KRAS-mut – KRAS-mutated; ECOG PS – performance status; ChT – chemotherapy; p – p value

Within normal range 17 5 12 Elevated (>37 UI/L) 32 15 17 Mean±Std.dev 3043±15554 998±2788 4453±20109 Median 136 98 230 (IQR) (1-108848) (1-12547) (1-108848)

TNM stage

I 12 6 6 II 13 4 9 III 11 6 5 IV 19 7 12

Clinical stage

Resectable 28 12 16 Borderline resectable 7 3 4 Locally advanced 2 1 1 Metastatic 18 7 11

Treatment

Surgery ± ChT 29 14 15 Only ChT (palliative) 20 6 14 Best supportive care 6 3 3 resectable disease, two were unfit for surgery and were also offered palliative ChT.

First-line adjuvant ChT regimens included gemcitabine monotherapy in 5 (28%) patients, gemcitabine plus capecitabine in 5 (28%) patients, FOLFIRINOX in 6 (33%) patients and capecitabine monotherapy in 2 (11%) patients. Of the 20 patients with locally advanced (unresectable) or metastatic disease, only 14 (70%) were treated with palliative ChT due to poor performance

P G12D G12V G12R G12C p

0.21 6 7 1 0 0.19 9 4 3 2 76 68 70 65 0.90 0.34 73±11 66±11 66±4 65±9 0.93 13 10 4 2 0.70 2 1 0 0 0.24 7 2 2 1 0.56 7 7 2 1 265±358 12957±35969 1890±2516 628±884 0.99 0.35 70 1120 1117 628 (1-1049) (1-108848) (5-5321) (3-1253) 0.59 5 1 0 0 0.28 2 4 2 1 4 1 0 0 4 5 2 1 0.99 9 5 1 1 0.43 3 0 1 0 0 1 0 0 3 5 2 1 0.40 8 5 1 1 0.74 6 4 3 1 1 2 0 0 status in the remaining six. These six patients had a rapid deterioration and received best supportive care only. First-line palliative ChT regimens included FOLFIRINOX in 8 (57%) patients, gemcitabine plus nab-paclitaxel in 4 (29%) patients and gemcitabine monotherapy in 2 (14%) patients.

Performance of ARMS/HRMA technique compared to sanger sequencing

Tumor samples: ARMS is based on the use of primers whose 3´

nucleotides are allele specific. Nevertheless, in the case of heterozygous samples (G12V/wt or G12D/wt), ARMS amplification as a stand-alone technique still shows amplification with both nucleotide specific primers (G12V mutated primer or G12D mutated primer) and WT primer. Additionally, the specificity and sensitivity of the ARMS reaction is dependent on the nucleotide that causes the mismatch; there are pairs of mismatches that cause higher destabilization of the polymerase, such as the one present in G12D mutation (GGT→GAT). On the contrary, for the G12V mutation (GGT→GTT) the destabilization caused by the mismatch is lower, meaning the polymerase still manages to amplify in the presence of WT samples. To avoid this type of issues, HRMA was immediately performed on the ARMS products (same tube without opening it) for an undoubtedly discrimination between genotypes, mainly the heterozygous G12V. In fact, HRMA is a more recent method that stands out for being fast and cost-effective for high-throughput mutation screening [22,23]. HRMA is based on the dissociation behaviour of DNA when subjected to increasing temperatures, in the presence of saturating fluorescent dyes with greater affinity for double-strand DNA than for single-strand DNA [22-24]. HRMA has been shown to detect between 3% to 10% of mutant DNA in a background of wild-type DNA [22]. In this regard, ARMS/HRMA was optimized to detect the most frequent mutations in KRAS (G12D and G12V). After excluding patients harbouring G12C and G12R mutations (6 out of 55 tumors) and 11 samples with no DNA available after SS, a population of 38 patients was used to compare SS and ARMS/HRMA techniques. Results of SS and ARMS/HRMA were successfully obtained from all the 38 tumor samples and are summarized in table 2. SS detected KRAS mutations in 23 (60%) patients: G12D in 13 and G12V in 10 patients. ARMS/HRMA was able to detect the same KRAS mutations detected by SS and an additional 10 mutated samples, with an overall of 33 KRAS mutated (87%) patients: G12D in 18, G12V in 12 and both G12V/G12D in 3 patients. There was a good agreement between the two techniques (kappa 0.215, p<0.001).

Liquid biopsies (Plasma samples): Regarding liquid biopsies, we were able to extract DNA from plasma samples to determine KRAS genotype by SS and ARMS/HRMA in 32 patients. As SS was not conclusive for 10 patients due to poor sequencing quality, only the results for the remaining 22 patients were further compared. SS was able to detect a single KRAS G12V mutation in 1 (4.5%) patient. ARMS/HRMA was successful in detecting KRAS mutation in the same patient (from SS) and in 6 additional patients (total of 32%): G12V in 3, G12D in 3 and G12V/G12D in 1 patient. Results are summarized in table 3. There was a good agreement between the two techniques (kappa 0.265, p<0.001).

Prognostic value of tumor KRAS mutations: Mean follow-up time of the overall population was 13±13 months (median 9 months, ranging from 0 to 51 months). In order to assess the prognostic value of KRAS mutations, we considered two cohorts of patients according to treatment performed, either surgical resection or palliative chemotherapy/best supportive care.

Surgical cohort: Considering only operated patients (Table 4), we observed that 49% of operated patients (14/29) had a WT tumor. Also, those with tumors harboring KRAS mutation had higher mortality (80% vs 36%, p=0.02) – Table 5. No significant differences were found regarding disease progression and/or according to different KRAS mutations.

In a survival analysis (Figure 3), KRAS-mutated patients had a lower PFS (median 5 vs 13 months, p<0.01), with a lower median PFS in patients harboring G12C, G12R and G12D mutations (G12C=0 vs G12R=4 vs G12D=4 vs G12V=9, p<0.01) – Figure 4. Similarly, KRAS-mutated patients had a lower OS (median 9 vs 14 months, p<0.01), with a lower median OS in patients harbor-

Table 5: Events during follow-up in the surgical cohort according to tumor KRAS genotype.

Abbreviations: KRAS-wt – KRAS wild-type; KRAS-mut – KRAS-mutated; p – p value. Results in N (%).

Total KRAS-wt KRAS-mut p Progression 11 4 (29%) 7 (47%) 0.32 Death 17 5 (36%) 12 (80%) 0.02 ing G12C, G12R and G12D mutations (G12C=0 vs G12R=4 vs G12D=9 vs G12V=11, p<0.01) – Figure 4.

Palliative and BSC cohort: Regarding the cohort of patients submitted to palliative ChT or BSC (Table 6), we observed that 35% (9/26) had WT tumors. Patients with KRAS mutated tumors tended to have less disease progression than patients who were KRAS wild-type (41% vs 78%, p=0.07) – Table 7. No significant differences were found regarding mortality rate or different KRAS mutations.

In a survival analysis, no significant differences were found in PFS and OS between KRAS-mutated and KRAS-WT, neither between different KRAS mutations.

Prognostic value of plasma KRAS mutations

Surgical cohort: In patients submitted to surgery (Table 8), those with ctDNA detectable KRAS mutation had a lower rate of disease progression (0% vs 50%, p=0.04). No significant differences were observed regarding mortality rate or between different KRAS mutations, although the two patients that died had a ctDNA G12D mutation (Table 9).

In a survival analysis, no significant differences were found in PFS and OS between patients with ctDNA KRAS mutation and KRAS- WT, neither between different KRAS mutations.

Palliative and BSC cohort: Regarding the cohort of patients treated with palliative ChT or BSC (Table 10), no significant differences were found in disease progression and mortality rates between patients with ctDNA KRAS mutation compared to KRAS-WT (Table 11). Similarly, no differences were found between different KRAS mutations.

In a survival analysis, patients who had ctDNA with KRAS mutation had a lower PFS (median 1 vs 6 months, p=0.02) and lower G12V G12R G12D G12C p 2 (40%) 0 (0%) 5 (63%) 0 (0%) 0.33 4 (80%) 1 (100%) 6 (75%) 1 (100%) 0.8 OS (median 1 vs 8 months, p<0.01) – Figure 5. When considering different KRAS mutations no significant differences in PFS and OS were observed between patients harboring G12D (median 1 month) or G12V (median 1 month) mutations; once again, patients with any of these two mutations had a tendency for a lower PFS (median 6 months, p=0.05) and a lower OS (median 8 months, p=0.02) – Figure 6.

Discussion

In the present study, we prospectively evaluated a group of 55 patients with newly diagnosed PDAC, to clarify the prognostic value of KRAS mutations. Furthermore, to determine the KRAS mutational status we applied and validated a technology based on ARMS coupled with HRMA and compared the results with SS.

We demonstrated that ARMS/ HRMA technique is more sensitive than the gold standard Sanger Sequencing to detect KRAS mutations both in tumours and plasma of patients with PDAC. Besides its higher sensitivity, ARMS/HRMA is a faster and cheaper technology, able to provide KRAS mutational status in a 6-hour period. Additionally, we observed that these mutations seem to have a prognostic value both in operated and chemotherapy treated patients.

The prognosis of PDAC remains very poor with only 9% of pa-

tients being alive at 5 years. Besides a late diagnosis in most of the cases, very little advances have been made in tailored therapies and all tumours are treated equally irrespective of their molecular characteristics. Currently, therapeutic decisions are based solely on cTNM staging despite previous studies showing the prognostic value of the type of KRAS mutations, an early and almost universal event in pancreatic carcinogenesis. Namely, KRAS G12D muta-

tion has been shown to be an independent predictor of worse survival whereas G12V mutations might identify those tumours that would respond better to chemo/radiotherapy [10–12]. One of the major reasons for molecular analysis not being considered in the therapeutic decisions of these patients, is certainly linked to the complexity and costs of molecular analysis which is incompatible with having available results in a short period of time for clinical decisions.

In this study, we validated a technology based on ARMS coupled with HRMA for a cohort of PDAC patients with stored tumor and plasma samples. ARMS/HRMA demonstrated to be a more sensitive technique with 100% specificity, being able to detect KRAS mutations in tumor samples in 10 additional patients (overall in 87% of patients) compared to SS (only detected KRAS mutations in 60%). This prevalence of KRAS mutations detected by ARMS/ HRMA is concordant to what is described in the literature using Next Generation Sequencing (NGS) or Droplet Digital PCR [5,6] but this technology is significantly faster and cheaper than NGS.

In regard to its prognostic value, we observed that in the surgical cohort, KRAS mutations in the tumor were associated with a lower PFS (median 5 vs 13 months, p<0.01) and lower OS (median 9 vs 14 months, p<0.01) compared to KRAS-WT; furthermore, G12C, G12R and G12D mutations were associated with lower survival times compared to G12V mutation (p<0.01). This discriminant value was not observed in chemotherapy treated patients.

In our perspective, these are important findings because at the present moment, although few PDAC patients go for direct surgery, this decision is mostly taken based on criteria of vascular invasion observed on CT/MRI. However, some patients, despite having resectable tumors, relapse shortly after. As pancreatic surgery is still an aggressive procedure with significant morbidity even in high volume centers, having the ability of selecting patients whose tumors have a higher probability of relapsing, could be additional decision criteria. In contrast, in the present study, tumors with KRAS mutations do not seem to have a worse prognosis when treated with chemotherapy. It will be important to confirm these observations in a larger sample of patients as well as try to understand whether specific mutations respond better to specific regimens of chemotherapies.

Liquid biopsy is an emerging technology that allows the noninvasive sampling of tumoral genetic material in circulation, and therefore is a promising tool to be used as a prognostic biomarker and to guide therapy [25]. Different types can be considered - circulating DNA, exosomes and circulating tumors cells. The latter are certainly the most sensitive but the more difficult to implement in clinical practice due to the complex and expensive technology involved. Although the sensitivity is much lower when using ctDNA, in a recent meta-analysis, both the presence of mutated ctDNA and higher concentrations of ctDNA were predictors of overall survival and progression free survival in PDAC [26]. Furthermore, ctDNA could potentially be used as a tool for disease monitoring since its detection is associated with the presence of micrometastases and could predict disease relapse before it becomes detectable by imaging [27,28].

Since it is well known that ctDNA has very low concentrations in plasma samples, a highly sensitive molecular analysis is needed to provide robust results [20]. In the present study, ARMS/HRMA

while SS had poor sequencing quality in 10 samples (69% success rate). Once again, ARMS/HRMA detected KRAS mutations in 7 patients (overall in 32% of patients) compared to SS which only detected KRAS mutations in one patient (4.5%). However, in our cohort the prevalence of KRAS mutation in ctDNA of PDAC patients is somewhat below to what has been reported in the literature (ranging from 44% to 67%) [29-32]. This can be due to the fact that we only analyzed 2 of the most frequent KRAS mutations; furthermore, different molecular technologies were used, as most studies used digital PCR or NGS, with no published reports of ARMS performance in this setting so far. Sample processing and storage might also be of paramount importance and could be further optimized. Despite the low prevalence of KRAS ctDNA mutation in our cohort detected by ARMS/HRMA, and considering only patients treated with chemotherapy, we observed that the presence of ctDNA harboring KRAS mutation at diagnosis was associated with lower PFS (median 1 vs 6 months, p=0.02) and lower OS (median 1 vs 8 months, p<0.01), with no significant differences being observed between G12D and G12V mutations. In the context of patients undergoing ChT, the use of liquid biopsies longitudinally could be particularly useful for assessing response to therapy. Our analysis was limited to ctDNA assessment only at diagnosis, but recently Kruger et al. have demonstrated that changes in KRAS-mutated ctDNA levels during ChT in 54 patients with advanced PDAC were more rapid and pronounced than traditional biomarkers. An early decrease in its levels was an indicator of early response to therapy and repeated measures during follow-up were superior to CA 19.9 in detecting disease progression [29]. Future studies should compare the performance of ARMS/ HRMA with other molecular technologies in plasma samples as well as exploring the association of different types of mutations In the surgical cohort, ctDNA detection showed no prognostic value which may be explained by our reduced sample size. In this context, a recent study including 112 patients has demonstrated that PDAC patients with detectable ctDNA harboring KRAS mutations pre- and/or post-surgery had lower recurrence-free survival and lower OS, with recurrence of disease occurring in all patients with ctDNA detectable after resection surgery [30]. These findings should be corroborated by further studies, so treatment intensification strategies with more effective ChT regimens could be adopted in the presence of detectable ctDNA.

Two major limitations of our study are the reduced sample size (even more reduced due to the number of samples with available DNA to be used simultaneously in SS and ARMS/HRMA) and the subsequent reduced number of events during follow-up, which may have limited the statistical power of the data analysis. To have an acceptable sample size and get robust results, a prospective multicentric study including more patients would be needed.

Conclusion

In conclusion, ARMS/HRMA technique seems to be a fast, accurate, cheaper and reliable diagnostic tool to assess KRAS mutations in PDAC tumor and plasma samples, with a higher sensitivity compared to SS. In our cohort, tumor and plasma codon 12 KRAS mutations in PDAC patients were associated with lower OS and PFS, more importantly the G12D mutation. The detection of KRAS mutations in tumor and plasma samples could help to guide therapeutic decisions in patients with PDAC. Future studies should corroborate our results regarding ARMS/HRMA and compare it to other molecular techniques, so that in the near future we could have a tool to access KRAS mutational status in PDAC patients readily available in routine clinical practice.

Acknowledgements

This work was supported by a grant from Grupo Luz Saúde, and is financed by national funds from FCT - Fundação para a Ciência e a Tecnologia, I.P., in the scope of the project UIDP/04378/2020 and UIDB/04378/2020 of the Research Unit on Applied Molecular Biosciences - UCIBIO and the project LA/P/0140/2020 of the Associate Laboratory Institute for Health and Bioeconomy - i4HB. FCT-MCTES is also acknowledged for 2020.07660.BD for BO.

References

1. GBD 2017 (2019) Pancreatic Cancer Collaborators. The global, regional, and national burden of pancreatic cancer and its attributable risk factors in 195 countries and territories, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. lancet Gastroenterol Hepatol, 4(12): 934-947.
2. Rahib, L., Smith, BD., Aizenberg, R., Rosenzweig, AB., Fleshman, JM., et al. (2014) Projecting Cancer Incidence and Deaths to 2030: The Unexpected Burden of Thyroid, Liver, and Pancreas Cancers in the United States. Cancer Res, 74(11): 2913-2921.
3. Rawla, P., Sunkara, T., Gaduputi, V. (2019) Epidemiology of Pancreatic Cancer: Global Trends, Etiology and Risk Factors. World J Oncol, 10(1): 10-27.
4. Khomiak, A., Brunner, M., Kordes, M., Lindblad, S., Miksch, RC., et al. (2020) Recent Discoveries of Diagnostic, Prognostic and Predictive Biomarkers for Pancreatic Cancer. Cancers (Basel), 12(11): 3234.
5. Giri, B., Sethi, V., Dudeja, V., Banerjee, S., Livingstone, A., et al. (2017) Genetics of pancreatic cyst-cancer progression. Curr Opin Gastroenterol, 33(5): 404-410.
6. Buscail, L., Bournet, B., Cordelier, P. (2020) Role of oncogenic KRAS in the diagnosis, prognosis and treatment of pancreatic cancer. Nat Rev Gastroenterol Hepatol, 17(3): 153-168.
7. Fan, Z., Fan, K., Yang, C., Huang, Q., Gong, Y., et al. (2018) Critical role of KRAS mutation in pancreatic ductal adenocarcinoma. Transl Cancer Res, 7(6): 1728-1736.
8. Huang, L., Guo, Z., Wang, F., Fu, L. (2021) KRAS mutation: from undruggable to druggable in cancer. Sig Transduct Target Ther, 6(1): 386.
9. Liberko, M., Kolostova, K., Szabo, A., Gurlich, R., Oliverius, M., et al. (2021) Circulating Tumor Cells, Circulating Tumor DNA and Other Blood-based Prognostic Scores in Pancreatic Ductal Adenocarcinoma – Mini-Review. In Vivo (Brooklyn), 35(1): 31-39.
10. Bournet, B., Muscari, F., Buscail, C., Assenat, E., Barthet, M., et al. (2016) KRAS G12D Mutation Subtype Is A Prognostic Factor for Advanced Pancreatic Adenocarcinoma. Clin Transl Gastroenterol, 7(3): E157.
11. Safi, S-A., Haeberle, L., Goering, W., Keitel, V., Fluegen, G., et al. (2022) Genetic Alterations Predict Long-Term Survival in Ductal Adenocarcinoma of the Pancreatic Head. Cancers (Basel), 14(3): 850.
12. Reddy, AV., Hill, CS., Sehgal, S., Ding, D., Hacker-Prietz, A., et al. (2021) Impact of somatic mutations on clinical and pathologic outcomes in borderline resectable and locally advanced pancreatic cancer treated with neoadjuvant chemotherapy and stereotactic body radiotherapy followed by surgical resection. Radiat Oncol J, 39(4): 304- 314.
13. Ako, S., Nouso, K., Kinugasa, H., Dohi, C., Matushita, H., et al. (2017) Utility of serum DNA as a marker for KRAS mutations in pancreatic cancer tissue. Pancreatology, 17(2): 285-290.
14. Cheng, H., Liu, C., Jiang, J., Luo, G., Lu, Y., et al. (2017) Analysis of ctDNA to predict prognosis and monitor treatment responses in metastatic pancreatic cancer patients. Int J Cancer, 140(10): 2344-2350.
15. Hadano, N., Murakami, Y., Uemura, K., Hashimoto, Y., Kondo. N., et al. (2016) Prognostic value of circulating tumour DNA in patients undergoing curative resection for pancreatic cancer. Br J Cancer, 115: 59-65.
16. Pietrasz, D., Pécuchet, N., Garlan, F., Didelot, A., Dubreuil, O., et al. (2017) Plasma Circulating Tumor DNA in Pancreatic Cancer Patients Is a Prognostic Marker. Clin cancer Res, 23(1): 116-123.
17. Watanabe, F., Suzuki, K., Tamaki, S., Abe, I., Endo, Y., et al. (2019) Longitudinal monitoring of KRAS-mutated circulating tumor DNA enables the prediction of prognosis and therapeutic responses in patients with pancreatic cancer. PLoS One, 14(12): e0227366.
18. Earl, J., Barreto, E., Castillo, ME., Fuentes, R., Rodrí- guez-Garrote, M., et al. (2021) Somatic Mutation Profiling in the Liquid Biopsy and Clinical Analysis of Hereditary and Familial Pancreatic Cancer Cases Reveals KRAS Negativity and a Longer Overall Survival. Cancers (Basel), 13(7): 1612.
19. Strickler, JH., Satake, H., Hollebecque, A., Sunakawa, Y., Tomasini, P., et al. (2022) First data for sotorasib in patients with pancreatic cancer with KRAS p.G12C mutation: A phase I/II study evaluating efficacy and safety. J Clin Oncol, 40(36): 360490-360490.
20. Sivapalan, L., Kocher, HM., Ross-Adams, H., Chelala, C. (2021) Molecular profiling of ctDNA in pancreatic cancer: Opportunities and challenges for clinical application. Pancreatology, 21(2): 363-378.
21. Bunduc, S., Gede, N., Váncsa, S., Lillik, V., Kiss, S., et al. (2021) Prognostic role of cell-free DNA biomarkers in pancreatic adenocarcinoma: A systematic review and meta-analysis. Crit Rev Oncol Hematol, 169.
22. Shackelford, RE., Whitling, NA., McNab, P., Japa, S., Coppola, D. (2012) KRAS Testing: A Tool for the Implementation of Personalized Medicine. Genes Cancer, 3(7-8): 459-466.
23. Reed, GH., Kent, JO., Wittwer, CT. (2007) High-resolution DNA melting analysis for simple and efficient molecular diagnostics. Pharmacogenomics, 8(6): 597-608.
24. Santos, S., Marques, V., Pires, M., Silveira, L., Oliveira, H., et al. (2012) High resolution melting: improvements in the genetic diagnosis of hypertrophic cardiomyopathy in a Portuguese cohort. BMC Med Genet, 13: 17.
25. Grunvald, MW., Jacobson, RA., Kuzel, TM., Pappas, SG., Masood, A. (2020) Current Status of Circulating Tumor DNA Liquid Biopsy in Pancreatic Cancer. Int J Mol Sci, 21(20): 7651.
26. Fang, Z., Meng, Q., Zhang, B., Shi, S., Liu, J., et al. (2021) Prognostic value of circulating tumor DNA in pancreatic cancer: a systematic review and meta-analysis. Aging (Albany NY), 13(2): 2031-2048.
27. Qi, Z-H., Xu, H-X., Zhang, S-R., Xu, JZ., Li, S., et al. (2018) The Significance of Liquid Biopsy in Pancreatic Cancer. J Cancer, 9(18): 3417-3426.
28. Guo, S., Shi, X., Shen, J., Gao, S., Wang, H., et al. (2020) Preoperative detection of KRAS G12D mutation in ctDNA is a powerful predictor for early recurrence of resectable PDAC patients. Br J Cancer, 122(6): 857-867.
29. Kruger, S., Heinemann, V., Ross, C., Diehl, F., Nagel, D., et al. (2018) Repeated mutKRAS ctDNA measurements represent a novel and promising tool for early response prediction and therapy monitoring in advanced pancreatic cancer. Ann Oncol Off J Eur Soc Med Oncol, 29(12): 2348-2355.
30. Lee, B., Lipton, L., Cohen, J., Tie, J., Javed, AA., et al. (2019) Circulating tumor DNA as a potential marker of adjuvant chemotherapy benefit following surgery for localized pancreatic cancer. Ann Oncol, 30(9): 1472-1478.
31. Bernard, V., Kim, DU., Anthony, F., Castillo, J., Allenson, K., et al. (2019) Circulating Nucleic Acids Are Associated With Outcomes of Patients With Pancreatic Cancer. Gastroenterology, 156(1): 108-118.e4
32. Patel, H., Okamura, R., Fanta, P., Patel, C., Lanman, RB., et al. (2019) Clinical correlates of blood-derived circulating tumor DNA in pancreatic cancer. J Hematol Oncol. 12(1): 1-12.