CRC is the second most commonly diagnosed cancer and a leading cause of death in Europe. Today, the median overall survival for patients (pts) with mCRC is ∼30 months. The treatment strategy normally includes a combination chemotherapy with or without biological treatment. The primary endpoint for our study was to describe retrospectively the disease's clinical pattern and treatment sequence pathway of mCRC patients who had received three lines of treatment in real-life by using natural language processing (NLP). We explored how NLP can extract and interpret clinical information from patients' electronic health records (EHRs). This is a multicentre, observational retrospective study based on the secondary use of 27,840,957 real-world data captured in the EHRs, of which, 1,521,128 were obtained from 1,608pts with mCRC who started 1st-line treatment using NLP and Artificial Intelligence techniques developed by Medsavana. 217 pts started 1st line and reached 3rd line during the study period (PAS population). A cross-sectional analysis was performed over patients who attended any of the 5 participating hospitals from 2013 to 2017. In PAS population, 135pts were male (62.2%) and the mean age was 63.7 years. By the time mCRC 1st line started, 31pts (14.3%) had ECOG reported, 29pts (96.3%) had ECOG 0-1 and 69pts (31.8%) showed primary tumour location at the left colon. 143pts (65,9%) had available in the eHCR their KRAS mutational profile, 110pts (50.7%) had NRAS status reported Microsatellite Instability and BRAF status were detected directly from free text in 25pts (11,5%) and 13pts (6%) respectively. Complete treatment sequencing across three lines was described for 217pts. 95 pts (43.8%) started the 1st line with a combination of Chemo+AntiVEGF, 99pts (45.6%) received Chemo+AntiVEGF in 2nd line and 61pts (28.1%) started 3rd line with chemo+AntiVEFG. For the combination of Chemo+antiEGFR, 50 pts (23%) started in 1st line, 32pts (14.7%) in 2nd line and 42pts (19.4%) in the 3rd line. Finally, there were 70pts (32%) who started 1st line with Chemo alone, 78pts (35.9%) in 2nd line and 84pts (38.7%) started 3rd line with Chemo alone. Through this NLP approach, clinical characteristics and treatment patterns in real world settings are consistent with previous published data in mCRC. To minimise the limitations caused by the data heterogeneity, higher levels of standardisation in data recording should be achieved. NLP represents a new exploratory approach to real world evidence studies. These Data represent a potential opportunity to explore the improvement of the clinical management of these patients.

CTCs have been extensively studied in different neoplasms. There are several techniques for their identification, however, it is not clear which is the most effective. In pancreatic cancer (PC) their significance in different stages of the disease and correlation with tumor markers are poorly understood. Prospective, observational, cohort study in two Spanish centers. We analyzed CTCs at day 0 (before chemotherapy). Isolation was carried out with Isoflux technology, combining cell separation by immunomagnetic particles (positive selection) with a microfluidic system. Four different markers were used, including two mesenchymal markers, EpCAM and EGFR. The count of CTCs was done in a confocal microscope using the iMSRC (intelligent Matrix Screen Remote Control). Primary endpoint was the correlation of CTCs with overall survival (OS). Secondary endpoints: correlation between CTCs with Ca19.9/CEA values and the difference in the median value of CTCs in metastatic compared to locally advanced disease were secondary endpoints. Sixty-three patients were analyzed. Clinical characteristics: (Table) Median follow up was 11.6 months. Median value of CTC-0 was 343 (range 33 - 6259) for metastatic patients while it was 436 (81 - 1082) for locally advanced patients (p = 0.380). Correlation coefficient for Ca 19.9 was 0.03 (p = 0.824); CEA correlation coefficient was 0.121 (p = 0.366). Kaplan-Meier analysis showed an estimated median OS of 9.9 months for > 500 CTCs (95% confident interval [95% CI] 8.1 – 11.6) vs 14.7 (95% confident interval [95% CI] 6.7 – 22.6) for < 500 CTCs (p = 0.021). Table: 1654PClinical characteristicsResultsMedian age (range) - yr65 (45 – 85)Sex – no. (%) Male Female35 (55.67) 28 (44.4)Stage at diagnosis – no. (%) Borderline Locally advanced Metastatic8 (12.7) 16 (25.4) 39 (61.9)BMI – no. (%) Underweight Normal Overweight Obesity2 (3.1) 34 (54) 21 (33.4) 6 (9.5)PS - no. (%) 0 1 212 (19) 43 (68.3) 8 (12.7)CEA – median (range)4.8 (0.5 – 381)Ca 19.9 – median (range)1131 (2 -6627800) Open table in a new tab With an increased number of patients there is still absence of differences of CTC-0 between metastatic and non-metastatic patients as well as no correlation of CTCs with tumor markers at diagnosis. 500 CTCs continues to be a significant cutoff for overall survival.

Cancer patients have a risk up to nine times higher of suffering from venous thromboembolism (VTE) compared to those without cancer, which makes VTE the second preventable cause of mortality in cancer patients. Pancreatic carcinoma is a high thrombotic risk tumor, being its incidence of thrombosis of up to 30% in outpatient patients. It has been suggested that circulating tumor cells (CTC) may be related to thrombosis, as they can alter blood coagulation and due to their pro-inflammatory and pro-coagulant properties. We conducted an observational, prospective, cohort study in a two Spanish centers. CTC were isolated from a blood sample of newly diagnosed patients of pancreas carcinoma before starting chemotherapy treatment. CTC count was determined by using Isoflux® technology followed by counting via confocal microscopy. Patients were followed up according to routine practice. Clinical and analytical data were collected up to death. We analyzed the relation between CTC and VTE, and between CTC and VTE scores. This analysis included a total of 63 patients newly diagnosed before chemotherapy was given. 39 (61,9%) patients were metastatic, 16 (25,4%) locally advanced and 8 (12,7%) borderline tumors. Median follow up was 11,63 months. 25 episodes of VTE occurred in 23 patients during follow up (11 pulmonary embolism, 5 deep vein thrombosis, 6 visceral thrombosis, 3 catheter-related thrombosis). Median value of CTC was 397 (range 33-21697). For a cut off of CTC 500, we found no relation between CTC and VTE (p=0,298). However, when we analyzed CTC>500 and VTE excluding visceral and catheter-related thrombosis, it was found to be associated (p=0.006, p=0.002 in those metastatic). We found no relation between CTC and thrombosis scores (Khorana, Viena, Protech, Conko nor CAT). This study shows that CTC detection before first-line chemotherapy could be a predictor marker for VTE.

During the first year of the SARS-CoV-2 pandemic the management and treatment of COVID-19 have been improved. However, cancer patients continue to be one of the most affected. We evaluate the mortality rate due to COVID-19 and associated risk factors in the cancer population diagnosed in our center during the first year of pandemic. We retrospectively reviewed the medical records of 189 cancer patients who were diagnosed with COVID-19 between March 5, 2020 and February 28, 2021. Mortality rate nd associated risk factors were studied. Mortality rate: 55/189 patients. Mean age: 72 years (34-95), 125/189 male patients. Predominant histologies: lung cancer (72/189), colorectal (31/189), breast (24/189). Predominant staging: metastatic disease (113/189). Predominant cancer treatment: chemotherapy (63/189); 118/189 patients were receiving any type of oncological treatment with palliative intention. Mortality was associated with male gender (45/55 vs 10/55, p=0.004), presence of comorbidities (48/55 vs 7/55, p=0.01), lung cancer (28/72 deaths with this tumor vs 27/117 with the rest, p=0.02), palliative intention cancer treatment (41/55 vs 12/55, p=0.02), older median age (76 vs 71, p = 0.02), higher median CRP (p=115.6 mg/dl vs 46 mg/dl), lower median lymphocytes (600/mm3 vs 1000/mm3 p<0.001). No specific treatment against COVID-19 significantly decreased mortality. Neither IL-6 nor ferritin were prognostic biomarkers. In multivariate analysis, male gender (OR 2.58, 95% CI 1.1-5.9, p = 0.02), lung cancer (OR 2.0, CI 1.0-3.8, p = 0.03), cancer treatment with palliative intention (OR 2.4, CI 1.07-5.3, p = 0.03), higher median CRP (OR 1.0, CI 1.00-1.01, p <0.001), as well as low lymphocyte median (OR 0.5, CI 0.25-1.0, p = 0.56), continued to be evidenced as risk factors, regardless of comorbidities, staging, sex, and palliative intention cancer-specific treatment, among other variables. Men with lung cancer under cancer-specific treatment with palliative intention who present, at the diagnosis of SARS-CoV-2 infection with elevated CRP above 115 mg/dl and a decrease in lymphocytes below 600/mm3 have a higher risk of presenting fatal complications.

656 Background: Randomized clinical trials have established new chemotherapeutic standards of care for metastatic pancreatic cancer, namely FOLFIRINOX (FFX) and gemcitabine + nab-paclitaxel (GNP) after demonstrating a significant and relevant increase of overall survival. However, there are some important uncertainties regarding how many patients are candidate to each of the two new regimens in the real life and how is the pattern of use in the elderly population. Methods: This is a retrospective study. Departments of Pharmacy of 7 Spanish hospitals generated the listings of patients (pts) treated in first line with these new regimens (FFX or GNP). Non-metastatic patients were excluded. An exploratory analysis was performed in the elderly population. Results: From Jan 2012 to Dec 2017, a total of 119 pts (M/F 58/42 %) were treated. Med age 63 y (38-83 y), 99% adenocarcinoma. 40% located in the head of pancreas. ECOG 87% 0-1. 89% had liver mets. In the 1st line 49.6% were treated with FFX and 50.4% with GNP. 53% of the pts could receive a 2nd line (82% after FFX 75% after GNP). The median OS was 12 months with no statistically significant differences between both regimens (12,7m for FFX vs 10,2 m for GNP). Elevated Ca 19.9 levels and Neutrophil-Lymphocyte ratio (NLR) increased the risk of death. Patients who received both regimens in first/second line had a median OS longer than 15 months whichever the sequence. 32 patients (27%) were older than 70 yo. 13 (41%) were treated with FFX and 19 (59%) with GNP. The median OS for patients older than 70 was 9.5m versus 12.3m for patients younger than 70. Conclusions: In our setting the use of FFX and GNP for treating metastatic pancreatic cancer is quite similar. Superiority could not be demonstrated for any of the schemes in first-line. Overall survival was determined by basal Ca 19.9 and NLR. Patients receiving both regimens (FFX or GNP) in first/second line whichever the sequence, exhibited the best survival rates. In our series elderly patients had poor survival rates.

Currently we still have limited information on how COVID-19 infection has affected lung cancer patients. In our study, we analysed whether there are differences in terms of mortality from COVID-19 between patients diagnosed with lung cancer and the overall population within our hospital health area (320,000 people). We have also studied the most frequent characteristics of lung cancer patients who develop infection with COVID-19, and we have analysed possible factors of poor prognosis, as well as treatment outcome. We performed a retrospective review of a total of 2216 patients admitted to Hospital Universitario Infanta Leonor in Madrid between March 5 and May 13, 2020 to identify the cumulative incidence of COVID-19 in patients with lung cancer and make a description of the characteristics of these patients, treatment outcome, risk factors for poor prognosis and mortality. We performed uni and multivariate logistic regression. 22/2216 of the total number of patients diagnosed with COVID-19 in our hospital had lung cancer (0.99%). 12/22 lung cancer patients with a COVID-19 diagnosis died (54.5%) vs 300/2216 COVID-19 patients in our hospital (p<0.0001). Lung cancer patients who died had a median age of 72 years (range of 49-84 years). Infection with COVID-19 in lung cancer patients was more frequent in men (72.73%). 18/22 (81.81%) had locally advanced or metastatic tumours. We observed a trend towards higher mortality among patients with hypertension than among non-hypertensive patients (10/15 vs 2/7; P=0.095). We found higher mortality among patients who developed acute respiratory distress syndrome (ARDS) than among those who did not (4/4 vs 8/12; P=0.044). There seems to be a trend towards lower mortality among patients who received treatment with the combination of hydroxychloroquine and azithromycin than among those who did not (6/14 vs 6/8; P=0.145). Lung cancer patients who became infected with COVID-19 have higher mortality than the general population. It is more frequent among men and the development of ARDS results in a worse prognosis with higher mortality. Although treatment with azithromycin and hydroxychloroquine appears to be a good treatment option, we must wait until we have more data on the safety of the combination and results in larger patient series.

Small case series of patients with active cancer and coronavirus infection have been described since the beginning of the pandemic. The patients most affected by this infection are those with lung cancer but it also affects other types of cancer such as breast cancer. We described the characteristics of patients with breast cancer and COVID 19, their associated risk factors, treatment and evolution. We reviewed 2216 medical records of all patients admitted to hospitalization with COVID-19 diagnosis between 5 March and 13 May 2020. Study data were collected and managed using RedCap electronic data capture tools. We described breast cancer patients, associated risk factors, mortality and outcome. We detected 85/2216patients cancer with a mortality rate 47% (40/85). Of all cancer patients, 11% (10/85) had breast cancer. Median age breast cancer patients was 70.5 years old (35-86). Most frecuent staging was locally advanced (50 %, 5/10) and most of them were on hormone therapy (50%, 5/10). As associated risk factors, 20% (2/10) had heart disease, 50% (5/10) had hypertension, 20 (2/10) were obese, 30% (3/10) had diabetes, 40% (4/10) had dyslipemia and only 10% (1/10) was smoker. Half the patients 50% (5/10) had bilateral pneumonia, none of them were admitted to the ICU and 20% (2/10) died. All patients were treated with the combination of azithromycin and hydroxychloroquine and 40% (4/10) with lopinavir/ritonavir. Mortality was associate with high LDH levels (1529 vs. 264 U/L, p=0,0002), high PCR levels (159.15 vs. 29 mg/L, p=0.0140), ARDS (1/1 vs. 1/9 without ARDS p=0.035). A posible relation has been found with history of hypertension (2/5 vs. 0/5 without hypertension, p=0.114) and bilateral pneumonia (2/5 vs. 0/5, p= 0.114). COVID 19 appears to have lower mortality in breast cancer patients than in other tumor types. High LDH and PCR levels and ARDS could be related with increased risk of death. Combined treatment in these patients with azithromycin and hydroxychloroquine might be a good option.

There are no large reported series determining the outcome of cancer patients with COVID-19. We aimed to determine whether differences exist in COVID-19 related mortality between cancer patients and the general population in our hospital, and we also describe associated risk factors. We reviewed 2216 medical records of all patients admitted to hospital with COVID-19 diagnosis between 5 March and 13 May 2020. Study data were collected using REDCap electronic data capture tools. We described COVID-19 cumulative incidence in cancer patients, treatment outcome, mortality and associated risk factors. We detected 85/2216 cancer patients in all COVID-19 diagnoses. Mortality rate: 40/85 cancer patients vs 260/2131 patients in the general ward (P<0.001). Median age: 76 years old (34-94), 50/85 male patients. Most frequent histologies: lung cancer (22/85), colorectal cancer (19/85) and prostate cancer (15/85). Most frequent staging: metastatic disease (32/85). Only 2/85 patients were admitted to ICU. Mortality was associated with older median age (79.5 vs 73, P=0.03), high d dimer levels (1630 vs 830, P=0.03), high LDH levels (315.5 vs 224, P=0.003), bilateral pneumonia (24/42 vs 5/22 with unilateral pneumonia, P=0.02), acute respiratory distress syndrome (ARDS) (12/13 vs 28/72 without ARDS, P<0.0001) and metastatic disease (20/32 metastatic vs 20/53 non-metastatic, P=0.02). Differences were maintained in multivariate analyses regarding ARDS (OR 23.7, P=0.007) and metastatic disease (OR 2.5, P=0.05). Combined treatment with hydroxychloroquine and azithromycin showed a better outcome in uni and multivariate analyses with only 21/61 dead patients (OR 0.13, P=0.005), adjusted by sex, histology, staging, ARDS and comorbidities. COVID-19 had significant mortality in cancer patients. High D dimer and LDH levels and ARDS development in elderly metastatic patients carry an elevated risk of death in cancer patients diagnosed with COVID-19. However, only 2/85 patients were admitted to ICU and this data was decisive. Combined hydroxychloroquine and azithromycin could be a good treatment option in COVID-19 cancer patients. It is a priority to create measures to avoid COVID-19 transmission in oncological patients.

Introduction: Nowadays, the use of FOLFIRI-aflibercept (FA) is approved in metastatic colorectal cancer patients as second-line chemotherapy after tumoral progression with an oxaliplatin-based regimen. Recent data suggest that resection of metastatic colorectal cancer metastases after second-line chemotherapy offers similar overall survival benefit compared to what has been described in first-line. The improvement in response rate with FA as second-line therapy may increase the number of patients who eventually undergo metastases resection. Methods: We performed an observational, retrospective analysis of 39 patients from 32 Spanish hospitals who underwent metastases resection after FA in a real-life setting. The objective of this analysis is to analyze the efficacy and safety (surgical morbidity and mortality) or this regimen. Results: The median age was 62.5 years old (42-83). 74.4% of the patients were male and 25.6%, female. 61.5 % presented ECOG 0 and 38.5%, ECOG 1. The primary tumors included were colon (71.8%) and rectum (28.2%). Tumor locations were: left 58.9%, right 30.8% and transverse 10.3%. RAS mutation was present in 71.8% of the tumors. The median follow up was 25.3 months (2.2-74.5). Most of the patients (68.4%) received a biological treatment combined with oxaliplatin-based chemotherapy: bevacizumab (n = 16, 42.1%), cetuximab or panitimumab (n = 9, 23.7%). All RAS wild-type tumors received anti-EGFR agents in first-line setting. Response rate to first-line chemotherapy: partial response (n = 16, 42.1%), progressive disease (n = 10, 26.3%), stabilization (n = 8, 21.1%), and complete response (n = 4, 10.5%). 37 patients (94.9%) received FOLFIRI-aflibercept as second line of treatment, with a median of 7 cycles prior to surgery (3-44). The best response to second line was partial response (n = 22; 56.4%), followed by stabilization (n = 15, 38.5%) and complete response (n = 2, 5.1%). Types of surgery performed: liver metastases (n = 18, 46.2%), pulmonary metastases (n = 9, 23.1%), peritoneal carcinomatosis (n = 6, 15.4%), others surgeries (n = 2, 5.1%) and simultaneous liver and peritoneal carcinomatosis (n = 4, 10.3%). Surgical results were the following: R0 (n = 28, 73.7%), R1 (n = 7, 18.4%) and R2 (n = 3, 7.9%). 20.5% of patients presented with postsurgical complications: abdominal abcess (25%), anastomosis deshicence (10%), biliary fistula (12.5%) and infectious diseases (50%). No patient presented with bleeding complications. The 30-90 days mortality rate was of 0%. 37.8% of the patients also received aflibercept post-surgery and a minor proportion (5.3%) required a second surgery. Median overall survival (mOS) after surgery (aSx) was 35.2 months (26.6-43.8). Median progression free survival (mPFS) aSx was 10 months (5.5-14.4). No difference in mOS between left and right side tumors was observed (P = .487). Conclusion: Resection of metastases after FA in oxaliplatin-refractory patients is safe and feasible with encouraging results in OS.

Introduction: Colorectal cancer (CRC) is the most incident in Spain, according to Spanish Network of Cancer Registries (REDECAN) and is the second cause of death by cancer in Spain. Although the OS in metastatic setting has increased to 24-30 months, the 5-year OS continues to be less than 12%. The first and second line of treatment are well defined; however, there are few therapeutic options for those patients who have progressed to these standard therapies. After progression to the second line of treatment, the only therapeutic options with approved specific indication are Regorafenib and Trifluridine and Tipiracil. Trifluridine and Tipiracil has shown its efficacy in two randomized clinical trials comparing Trifluridine and Tipiracil versus placebo: a Phase II and a Phase III study (RECOURSE). In the previous Phase II study, the mOS was 9.0 months in the group treated with Trifluridine and Tipiracil, compared to 6.6 months in the placebo group. In the RECOURSE study, among patients treated with Trifluridine and Tipiracil mPFS was 2 months versus 1.7 months among those treated with placebo, and mOS was 10,5 with Trifluridine and Tipiracil vs 7,6 with placebo. We described the clinical characteristic of patients treated with Trifluridine and Tipiracil in seven hospitals from Madrid; identifying and analyzing clinical factors associated with long-term response. Methods: We collected retrospectively the clinical data of 98 patients who had received treatment with Trifluridine and Tipiracil until January 2018 in seven different hospitals in Madrid. Results: The mean age at first use of Trifluridine and Tipiracil was 66± 9.45 years, 57.5% were men and 42.3% were women, most of them in good performance status (ECOG 0-1: 71.2%), 27.8% had ECOG ≥2 when started Trifluridine and Tipiracil. All of the patients were diagnosed of metastatic colorectal cancer (73.2% had liver metastases, 58.2% lung metastases and 30.9% peritoneal metastases). Tumor localization was: 73.2% left colon, 24.8% right colon, and only 2% small intestine or 1% two synchronous colonic tumors. KRAS mutations were found in 62.9%, NRAS mutations in 17.4% and BRAF mutations were found in 5.4% patients. The median of prior lines of chemotherapy was three. The median progression free survival was 3 months (95% CI 2.65 – 3.36). The median overall survival was 5 months (95% CI 4.23-5.78). The median duration of treatment was 3 cycles. 23.7% patients had an adverse event that led to treatment withdrawal. The requirement of dose-reduction was associated with longer PFS (4 months vs 3 months, p = 0.002) and OS (14 months vs 5 months, p = 0.017). The existence of BRAF mutation was also associated with shorter PFS (1 month vs 3 months, p = 0.002) and OS (1 month vs 5 months, p = 0.000). There was no statistically significant difference of PFS and OS according to primary tumor location. Conclusion: Trifluridine and Tipiracil is effective in metastatic colorectal cancer (in patients with ≥2 lines of treatment). The OS (5 months) and PFS (3 months) reached in our study in real clinical practice are consistent with findings from previous studies. Mutational status of BRAF was statistically significant associated with shortened PFS and OS. Dose reduction during treatment is associated with Trifluridine and Tipiracil efficacy in terms of prolonged OS and PFS.