The main objective of this amendment was to update the lurbinectedin dose from 4.0 mg/m^2 to 3.2 mg/m^2 based on new data available. This reduction of the starting dose avoided dose adjustments in patients with ECOG PS=2 and/or aged > 70 years, andmandatory CSF prophylaxis was removed. This amendment was implemented previous to treat any patient; therefore, all patients included in this study were treated with lurbinectedin 3.2 mg/m^2. This amendment also included the following modifications in the protocol: • The timing of the assessment of AAGP levels was clarified by removing it from the Biochemistry B list of assessments and including it as a separate item in the Schedule of Assessments. • The timing of the blood sample collection for the PGt analysis was modified. The sample was then collected at any time during the study, but preferably just before treatment start in Cycle 1. • The primary prophylactic antiemetics that need to be administered to the patients and their routes were clarified. • Version 1.1 of RECIST was implemented in 2009 and was in common use ever since. Thus, the table describing the differences between RECIST 1.0 and 1.1 was no longer required and was removed. • Study dates and contact information were updated.

The following changes were implemented in this amendment: • An ongoing clinical trial with lurbinectedin as single agent in patients with BRCA1/2-associated metastatic breast cancer previously untreated with poly (ADP-ribose) polymerase (PARP) inhibitors was amended to include also patients who have received prior therapy with PARP inhibitors. Therefore, no more patients with this disease should be included in this phase II Basket study, and all information on breast cancer was removed. • Some eligibility criteria were revised: • The types of H&N and GCTs that could be included into the study were clarified. • The classification of NETs was updated according to Bosman, F.T. et al. WHO Classification of Tumours of the Digestive System (IARC Press, Lyon, France, 2010). • The prior treatment requirement for NETs and biliary tract carcinoma was changed to allow the inclusion of patients after one or two prior chemotherapycontaining lines. • MRI was added as a valid method for detecting brain metastases in SCLC patients at baseline. • Fertile patients who did not use an effective method of contraception were excluded from participation in the study. • The criteria for treatment continuation was made consistent with the protocol’s eligibility criteria and the guidelines described in the Investigator’s Brochure for lurbinectedin, following a request by the UK’s Medicines and Healthcare Products Regulatory Agency (MHRA). Re-treatment of patients with ≤ 1.5 x ULN or creatinine clearance ≥ 30 mL/min was allowed. • Information on the statistical power of the study calculated using an exact binomial distribution was added. • The sample size of the cohort of SCLC patients in this study was to be increased to 50 evaluable patients if the success boundary (≥2 confirmed responses) was reached in the first 25 evaluable patients. This was done to collect further information on the efficacy and safety of single-agent lurbinectedin in this indication.

The protocol was amended to correct a typographic erratum in the inclusion criterion #10 describing the time when pregnancies must be avoided during the trial. In the previous substantial amendment #2, this time had already been changed from “six weeks after the last lurbinectedin administration” for all patients, to “three and four months after the last lurbinectedin administration for female patients and for partners of male patients, respectively” in other parts of the protocol. This amendment corrected this omission. Furthermore, a minor style edit change was also added (i.e., “men” instead of “male patients”).

The objective of this amendment was to allow up to 50 patients to be included in the endometrial carcinoma cohort of the study. Initially, up to 25 patients were planned per cohort to establish antitumor activity. In a previous amendment, the cohort of SCLC was increased to up to 50 patients (see Section 9.8.2). The current amendment allowed recruitment of up to 50 patients also in the endometrial carcinoma cohort. These two cohorts were expanded because of the antitumor activity already seen in a previous study of lurbinectedin in combination with doxorubicin (PM1183-A-003-10) and the need to confirm the activity of this compound as single agent in these indications. This change also affected the expected total number of patients. In response to Investigators’ input and in order to broaden the eligible patient population in the EFT cohort, the prior treatment requirement (#4h) was also changed to “no more than two prior chemotherapy-containing lines in the metastatic/recurrent setting”. The following changes, corrections, and clarifications were also included in this amendment: • PK sample collection windows were extended in response to feedback from study centers and to facilitate compliance with study procedures. • Drug-drug interaction information was updated in line with current PK analysis and the current Investigator’s Brochure (version 8.0, 10 March 2016). • In line with existing exclusion criterion #3, MRI was included as an option for baseline radiological tumor assessment in SCLC patients. • The criteria for treatment continuation included absence of active infection (including sepsis) and/or bleeding (any grade) in response to a requirement from the Belgian Competent Authority. • The definitions of ‘Day 0’ and associated assessment windows were clarified in order to ensure that patients had appropriate laboratory tests if first infusion was delayed. • Study contact details were updated.

The following changes were implemented in this amendment: • To allow up to 100 patients to be included in the SCLC cohort. Due to the results in advanced SCLC observed in a prior clinical trial of lurbinectedin in combination with doxorubicin as a second line (response rate: 50%; 95%CI, 34– 66%), the cohort of SCLC was expanded to include 50 patients in a previous amendment. The rationale to further increase the patient population of this cohort to 100 patients was to confirm the activity of lurbinectedin as a single agent in patients with SCLC, which was found so far in the PM1183-B-005-14 trial, and to support the ongoing phase III clinical trial of lurbinectedin in combination with doxorubicin in advanced SCLC (PM1183-C-003-14). The statistical methods section was amended to update the sequential test methodology and to provide further details on the control of type I and II error probability (alpha and beta), taking into account the two planned interim analysis performed at 15 and 25 patients per group for the two expanded cohorts, endometrial carcinoma (50 patients) and SCLC (up to 100 patients added to the protocol). In addition, for all 100 patients in the SCLC cohort, anonymized copies of tumor assessments (CT-scan or MRI) were requested to the investigational sites for a possible independent review. These patients were to be followed up until death to obtain survival results (in this cohort, the secondary endpoint was changed to overall survival instead of one-year overall survival). • The cohort of patients with metastatic breast cancer (MBC) positive for the germline mutations BRCA1 and BRCA2 was re-opened. MBC with BRCA1/2 cohort was closed, due to redundancy with another ongoing clinical trial. The rationale for re-opening of the cohort was, first, the activity of lurbinectedin seen in the patient population of MBC BRCA+already included in the study, leading to the request of patient inclusion by the Investigators participating in the trial.

• Study objectives. Assessment of antitumor activity by an IRC was included as a secondary objective in the SCLC cohort. The inclusion of this secondary objective in the new version of the protocol led to changes throughout different sections of the protocol (e.g., statistical analysis, secondary endpoints, etc.). In addition, the SAP was also updated accordingly. The submission of anonymized copies of tumor assessments CT-scan or MRI) from the investigational sites for a possible independent review was already implemented. These copies of CT scans, MRIs and any other documented methods to evaluate tumor response or progression in the SCLC cohort should be available for external radiological review by the IRC. • Patient population. The maximum number of evaluable patients was increased to 350 because of the elevated recruitment in the endometrial carcinoma cohort, which exceeded in 20 patients the planned number, together with the increase in the SCLC cohort to 100 evaluable patients, and the re-opening of the metastatic breast carcinoma BRCA-positive cohort, which had to recruit up to 25 patients. • Duration of recruitment period and total duration of the study. The duration of the recruitment period has been extended to approximately 40 months and the total duration of the study has been prolonged to 52 months. • Duration of follow-up period. The survival follow-up of thepatients within each individual cohort, except those of the SCLC cohort, will be the following: after documentation of progressive disease (PD) or start of a new therapy, patients will be followed-up every six months until death or until the date of study termination or clinical cutoff (i.e., when all evaluable patients of each cohort, have at least 12 months of follow-up from the first PM01183 infusion). Patients in the SCLC cohort, after PD will be followed-up every six months until death. • Overall survival (OS). OS rate at 6 months and 12 months will be determined in each cohort.

The objective of this amendment was to change the clinical cutoff and duration of follow-up of patients treated with lurbinectedin in the SCLC cohort because of data for the primary endpoint (ORR) were considered mature at the time of this amendment. A previous amendment established that patients in the SCLC cohort were to be followed-up until death to assure the recording of survival results (in this cohort, the secondary endpoint was changed to overall survival instead of one-year overall survival). In the SCLC cohort, the first lurbinectedin dose in the first recruited patient was administered on 27 October 2015, and the first lurbinectedin dose in the last recruited patient was administered on 16 October 2018. Hence, by the end of July 2020, more than 18 months of follow-up could have been collected for all patients recruited in this cohort. Taking into account this information, median OS in the SCLC cohort could be considered steady and no changes in the point estimates were foreseen. Therefore, this amendment defined the duration of follow-up of patients in the SCLC cohort as at least 18 months from the first lurbinectedin infusion. The changes implemented in this amendment could not affect the results of the study.