key: cord-0815776-b26a5lhd
authors: Fernández-Mañas, L.; Gorría, T.; Auclin, E.; Reyes, R.; Teixidó, C.; Marco-Hernández, J.; Padrosa, J.; De Herreros, M. García; Pesántez, D.; Martínez, D.; Mollà, M.; Vollmer, I.; Marrades, R.; Guirao, A.; Prat, A.; Viñolas, N.; Mezquita, L.; Reguart, N.
title: P09.15 Severity of Lung Cancer Disease in Hospitalized Patients During COVID-19
date: 2021-03-31
journal: Journal of Thoracic Oncology
DOI: 10.1016/j.jtho.2021.01.443
sha: 93de3ece14c0ef8e9f17a04ba41b9112aae319b0
doc_id: 815776
cord_uid: b26a5lhd

nan

Introduction: Worldwide evidence demonstrates that in metastatic NSCLC with an EGFR sensitizing mutation, EGFR inhibitors are superior to cytotoxic chemotherapy in the frontline treatment (1), However, in some countries limited resources force oncologists to start with chemotherapy until EGFR mutation test are available.

There are not data about the oncologic otucomes realted to this behavior in latam. Methods: An experimental natural study was conducted, using the natural situation in Peru, in which some patients start first-line cytotoxic chemotherapy before ITK due to poor access in obtaining the results for the EGFR gene mutation status. To do this, a real-world cohort was divided into two groups: Group 1: Patients with delayed molecular testing who started chemotherapy (duplet based on platins) and subsequent switch to EGFR TKIs and Group 2: Patients with timely results of mutational status for EGFR gene that started EGFR TKIs in the first line (erlotinib). For these analysis we developed a predictive methodology, using neural network models and COX regression models in which we make predictions as a function of time. Models were run in Orange ® software. Results: A total of 81 patients were collected during 5 years of follow-up. 62.9% corresponded to group No 1 and 37.1% to group 2. According to each characteristic, the model was able to predict the probability of survival or death with an accuracy of 92% in the living and 74% in the deceased. (Table 1 ). The Cox regression predictive model showed significant OR for group 2 (OR 0.28 p ¼ 0.013), and negative for OS with the INS exon 20 mutation (OR 28.29 p ¼ 0.00009) and L681Q mutation (OR 23.4 p ¼ 0.008). At the end of the model, it allowed predicting the probability of survival as a function of time and according to the characteristics of each patient (Table 1) .

The model worked to predict individualized OS, and we can adjust its probability according to time. As relevant data from the model, it was found that, for this real-world cohort, first-line treatment with chemotherapy and subsequent switch to EGFR ITKs in this case erlotinib (group 2) obtained a lower probability expressed in Odds for death. Likewise, the EGFR mutation subtype, specifically L681Q or INS exon 20, are the variable with the highest predictive weight of death in our population. This predictive model can be used in the clinic to make personalized recommendations adjusted to the natural settings of countries with similar access environments. Keywords: EGFR, neural network, non-small cell lung cancer Introduction: Covid-19 pandemic has drastically changed the management of patients with cancer; however, there is still limited data regarding the real impact of Covid-19 on patient's outcomes due to delayed diagnosis and treatment of clinical complications. We aimed to assess the prevalence, severity and mortality of clinical complications and oncology emergencies in hospitalized patients in our institution during the Covid19 period vs. the same period of 2019. Methods: We conducted a retrospective study of patients with small cell lung cancer (SCLC) and nonsmall cell lung cancer (NSCLC) who were admitted to the Department of Medical Oncology during Jan-Jun 2019 (before-Covid) and Jan-Jun 2020 (Covid-19 period). Clinical, pathological and biological data were collected. We assessed the clinical severity in both periods including: PS at admission, progression disease (PD), oncologic emergencies (%), start of a systemic therapy or switch to other therapy line. We also analyzed the differences on the 30-day mortality rate since hospitalization between both periods. Results: 229 admissions, 133 during and 93 before Covid-19 pandemic (N¼180 patients) were enrolled; the median duration of the hospitalization was 9 days (4-16). Median age was 66 years, 35% were female, 88% with PS2, 27% were current smokers; 83% had NSCLC histology. Most of them (82%) had advanced disease at admission; 69% were under systemic therapy (chemotherapy 39%, immunotherapy 17%, targeted therapies 11%). Nine patients (4%) were active covid-19 cases (9 NSCLC, 0 SCLC). The table 1 summarized the most common clinical conditions by histology, in both periods. In NSCLC population, during-Covid, lower rate of admissions were observed (4 cases less per month), with no increase of oncologic emergencies. The PD during hospitalization was slightly higher during vs. before-Covid, but no differences were observed in 30-days mortality rate. In SCLC population, during-Covid, the rate of admissions was doubled (2 cases more per month), with more cases progressing during the hospitalization. (46% during vs. 34% before-Covid). In contrast to NSCLC, the 30-days mortality rate was higher during-Covid (38%) vs. before-Covid (20%). Updated data will be presented in the meeting. 

Conclusion: We preliminary observed more aggressive disease with worse outcomes in patients with SCLC hospitalized during-Covid compared to the same period in 2019. No differences were observed in NSCLC. The final outcomes will be assessed in a larger and mature cohort still ongoing

Real-Life Experience from Castile and Leon S. Medina Valdivieso, 1 C. Bayona Antón, 2 L

De Frutos González, 2 R

Diz Taín 1 1 Medical Oncology

77.8% were males, 6.5% non-smokers and 71.8% stage IV; 14,4% had ECOG0, 58.6% ECOG1 and 26.7% had ECOG2; 70.5% had more than one comorbidity. Histological subtype was 55.4% adenocarcinoma, 38.5% squamous-cell and 6.1% NOS; PDL-1 was reported in 182 (38.1%) patients. Of the total, 85.1% received Nivolumab, 9.8% atezolizumab and 5% pembrolizumab; and 28% received 2 lines before ICI. Immune-related adverse events (IRAEs) were present in 26.4%; 75.7% were grade 1-2. The most frequent IRAEs were hypothyroidism (7%) and pneumonitis (7%)

NSCLC (N¼ 190) SCLC (N¼ 39)

N¼83 (44%) Before-Covid (2019) N¼107 (52%) During-Covid (2020) N¼13 (33%) Before-Covid (2019) N¼26 (67%) Others 27 (33%) 20 (24%) 9 (11%) 11 (13%) 16 (19%) 26 (24%) 22 (21%) 16 (15%) 20 (17%)

8%) 7 (8%) 69 (84%) 7 (7%

Sites at admission £2 >2 55 (67%) 28 (33%) 64 (60%) 43 (40%) 8 (62%) 5 (38%) 14 (54%)

Progression during hospitalization Yes No 32 (39%) 51 (61%) 35 (32%) 72 (68%) 6 (46%) 7 (54%)

Early death (30-days mortality rate) Yes No 20 (24%) 63 (76%) 22 (21%)