A Study to Evaluate the Safety and Efficacy of Polatuzumab Vedotin in Combination With Rituximab, Gemcitabine and Oxaliplatin Compared to Rituximab, Gemcitabine and Oxaliplatin Alone in Participants With Relapsed or Refractory Diffuse Large B-Cell Lymphoma

PFS=time from randomization to the first occurrence of disease progression (PD) (based on either: PET-CT data/not including any PET data), as determined by investigator, per Lugano response criteria or death due to any cause, whichever occurs first. PD based on PET-CT data=score 4 (uptake moderately > liver) or 5 (uptake markedly higher than liver \&/or new lesions) on 5-point scale (5PS) with an increase in intensity of uptake from baseline at individual target nodes/nodal masses and/or new FDG-avid foci extranodal lesions consistent with lymphoma at interim or end-of-treatment assessment. New lesions: New FDG-avid foci consistent with lymphoma rather than another etiology, Bone marrow: New/recurrent FDG-avid foci. Participants who did not report PD nor died at time of analysis were censored on the date of last evaluable tumor assessment if post-baseline tumor assessment or on date of randomization if no post-baseline tumor assessment. KM methodology was used to estimate median PFS.

CRR was defined as the percentage of participants who had a complete metabolic response (CMR) based on positron emission tomography-computed tomography (PET-CT) according to Lugano response criteria at the end of treatment as determined by IRC. CMR was defined as score 1, 2, or 3 (1=no uptake above background; 2=uptake ≤ mediastinum; 3=uptake >mediastinum but ≤ liver) with or without a residual mass on 5PS at lymph nodes and extra lymphatic sites. In Waldeyer's ring or extranodal sites with high physiologic uptake or with activation within spleen or marrow (e.g., with chemotherapy/myeloid colony-stimulating factors), uptake may be greater than normal mediastinum and/or liver. In this circumstance, CMR was inferred if uptake at sites of initial involvement was no greater than surrounding normal tissue even if the tissue had high physiologic uptake; no new lesions and no evidence of fluorodeoxyglucose (FDG)-avid disease in bone marrow. Percentages are rounded off.

ORR=percentage of participants with CMR/partial metabolic responses (PMR), based on PET-CT per Lugano response criteria as per an IRC. CMR=score 1, 2, or 3 with/ without a residual mass on 5PS at lymph nodes (LN) \& extra lymphatic sites. In Waldeyer's ring or extra nodal sites with high physiologic uptake or with activation within spleen/marrow, uptake may be greater than normal mediastinum \&/or liver. In this case, CMR was inferred if uptake at sites of initial involvement was no greater than surrounding normal tissue even if the tissue had high physiologic uptake; no new lesions \& no evidence of FDG-avid disease in bone marrow. PMR=score 4 or 5 on 5PS (4=uptake moderately > liver; 5=uptake markedly higher than liver \&/or new lesions) with reduced uptake compared with baseline and residual mass(es) of any size at LN \& extra lymphatic sites; no new lesions \& residual uptake higher than uptake in normal marrow but reduced compared with baseline. Percentages are rounded off.

BOR was defined as the best response while on study (based on PET-CT or CT data) according to Lugano response criteria, as determined by the investigator. CMR and PMR based on PET-CT data were defined as outlined in the ORR outcome measure (OM) number 8. CT-based complete radiologic response was defined as target lymphatic nodes/nodal masses regression to ≤ 1.5 centimeters (cm) in longest transverse diameter of a lesion (LDi); no extralymphatic sites of disease; absence of non-measured lesion; enlarged organs regress to normal; no new lesions and bone marrow normal by morphology, if indeterminate, immunohistochemistry (IHC) negative. CT-based partial response was defined as ≥50% decrease in sum of the product of the perpendicular diameters for multiple lesions (SPD) of up to 6 target measurable nodes and extra nodal sites; absent/normal, regressed, but no increase in non-measured lesions; spleen regressed by >50% in length beyond normal and no new lesions. Percentages are rounded off.

CRR was defined as the percentage of participants who had CMR based positron emission tomography-computed tomography (PET-CT) according to Lugano response criteria at the end of treatment as determined by investigator. CMR was defined as score 1, 2, or 3 (1=no uptake above background; 2=uptake ≤ mediastinum; 3=uptake >mediastinum but ≤ liver) with or without a residual mass on 5PS at lymph nodes and extra lymphatic sites. In Waldeyer's ring or extranodal sites with high physiologic uptake or with activation within spleen or marrow (e.g., with chemotherapy/myeloid colony-stimulating factors), uptake may be greater than normal mediastinum and/or liver. In this circumstance, CMR was inferred if uptake at sites of initial involvement is no greater than surrounding normal tissue even if the tissue has high physiologic uptake; no new lesions and no evidence of FDG-avid disease in bone marrow. Percentages are rounded off.

ORR=percentage of participants with CMR/PMR, based on PET-CT as determined by investigator per Lugano response criteria. CMR=score 1, 2, or 3 with/ without a residual mass on 5PS at LN \& extra lymphatic sites. In Waldeyer's ring or extra nodal sites with high physiologic uptake or with activation within spleen/marrow, uptake may be greater than normal mediastinum \&/or liver. In this case, CMR was inferred if uptake at sites of initial involvement was no greater than surrounding normal tissue even if the tissue had high physiologic uptake; no new lesions \& no evidence of FDG-avid disease in bone marrow. PMR=score 4 or 5 on 5PS (4=uptake moderately > liver; 5=uptake markedly higher than liver \&/or new lesions) with reduced uptake compared with baseline and residual mass(es) of any size at LN \& extra lymphatic sites; no new lesions \& residual uptake higher than uptake in normal marrow but reduced compared with baseline. Percentages are rounded off.

DOR=time from the date of the first occurrence of a documented objective response (complete or partial response [CR/PR]) (based on PET-CT or CT data) as determined by the investigator, using Lugano response criteria, until PD (based on either response: including PET-CT data or not including any PET data) or death, whichever occurred first. CMR and PMR based on PET-CT data were defined as outlined in the ORR OM. CT-based complete and partial response were defined as outlined in the BOR OM. PD defined as outlined in the PFS OM.

EFSeff was defined as the time from randomization to first to the earliest occurrence of the following: PD or relapse using Lugano response criteria (based on either response: including PET-CT data or not including any PET data); death due to any cause or initiation of any NALT. PD based on PET-CT data was defined as score 4 or 5 on 5PS with an increase in intensity of uptake from baseline at individual target nodes/nodal masses and/or new FDG-avid foci extranodal lesions consistent with lymphoma at interim or end-of-treatment assessment. New lesions: New FDG-avid foci consistent with lymphoma rather than another etiology, Bone marrow: New or recurrent FDG-avid foci. Participants with no EFSeff events were censored at the time of the last evaluable tumor assessment if post-baseline assessments were available and participants who did not undergo a post-baseline tumor assessment were censored at the time of randomization. KM method was used to estimate median EFS for each treatment arm.

Time to deterioration was defined as the time from randomization to the first documentation of a 10-point decrease in EORTC QLQ-C30 physical functioning scale. EORTC QLQ-C30 consists of 30 questions assessing five aspects of participant functioning (physical, emotional, role, cognitive, and social), three symptom scales (fatigue, nausea and vomiting, pain), global health status and quality of life (GHS/QoL), and six single items (dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial difficulties). The functioning items were scored on a 4-point scale that ranged from "not at all" to "very much". Raw average scale scores were linearly transformed to range 0-100 with higher scores indicating higher response levels (i.e. higher functioning). KM method was used to estimate median time to deterioration for each treatment arm.

Time to deterioration was defined as the time from randomization to the first documentation of a 10-point increase from baseline in EORTC QLQ-C30 fatigue scale. EORTC QLQ-C30 consists of 30 questions assessing five aspects of participant functioning (physical, emotional, role, cognitive, and social), three symptom scales (fatigue, nausea and vomiting, pain), GHS/QoL, and six single items (dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial difficulties). The symptom scale (fatigue) was scored on a 4-point scale that ranged from "not at all" to "very much". Raw average scale scores were linearly transformed to range 0-100 with higher scores indicating higher response levels (i.e. higher symptom severity). KM method was used to estimate median time to deterioration for each treatment arm.

Time to deterioration was defined as the time from randomization to the first documentation of a >3-point decrease from baseline in FACT-Lym LymS. The FACT-Lym subscale has 4 domains: physical well-being, social/family well-being, emotional well-being, functional well-being and lymphoma-specific subscale (LymS). The FACT-Lym LYMS consists of common lymphoma disease and/or treatment-related symptoms (e.g., pain, fever, swelling, night sweats, insomnia, itching, weight loss, fatigue, and loss of appetite). FACT-Lym LYMS contains 15 items scored from 0-4, where 0="Not at all" and 4="very much". Scale score range is from 0 to 60. Higher score indicates better quality of life. KM method was used to estimate the median time to deterioration for each treatment arm.

EORTC QLQ-C30 consists of 30 questions assessing five aspects of participant functioning (physical, emotional, role, cognitive, and social), three symptom scales (fatigue, nausea and vomiting, pain), GHS/QoL, and six single items (dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial difficulties). The functioning items were scored on a 4-point scale that ranged from "not at all" to "very much". Raw average scale scores were linearly transformed to range 0-100 with higher scores indicating higher response levels (i.e. higher physical functioning).

EORTC QLQ-C30 consists of 30 questions assessing five aspects of participant functioning (physical, emotional, role, cognitive, and social), three symptom scales (fatigue, nausea and vomiting, pain), GHS/QoL, and six single items (dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial difficulties). The symptom scale (fatigue) were scored on a 4-point scale that ranged from "not at all" to "very much". Raw average scale scores were linearly transformed to range 0-100 with higher scores indicating higher response levels (i.e. higher symptom severity). A positive change from baseline indicates an improvement.

The FACT-Lym subscale has 4 domains: physical well-being, social/family well-being, emotional well-being, functional well-being and lymphoma-specific subscale (LymS). The FACT-Lym LYMS consists of common lymphoma disease and/or treatment-related symptoms (e.g., pain, fever, swelling, night sweats, insomnia, itching, weight loss, fatigue, and loss of appetite). FACT-Lym LYMS contains 15 items scored from 0-4, where 0="Not at all" and 4="very much". Scale score range is from 0 to 60. Higher score indicates better quality of life. A positive change from baseline indicates an improvement.

The FACT/GOG-NTX scale provides a targeted assessment of symptoms of peripheral neuropathy, including sensory, motor, and auditory problems and cold sensitivity. It contains 12 items scored from 0-4, where 0="Not at all" and 4="very much". Scores are summed for a range of 12-44, with lower scores indicating more severe neurotoxicity. Participants completed assessments until progression or non-protocol-specified anti-lymphoma treatment (NALT). Only records on or before the first NALT were summarized.

EORTC QLQ-C30 consists of 30 questions assessing five aspects of participant functioning (physical, emotional, role, cognitive, and social), three symptom scales (fatigue, nausea and vomiting, pain), GHS/QoL, and six single items (dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial difficulties). The functioning items were scored on a 4-point scale that ranged from "not at all" to "very much". Raw average scale scores were linearly transformed to range 0-100 with higher scores indicating higher response levels (i.e. higher functioning). Clinically meaningful improvement was defined as a participant with at least a 7 point scale score increase.

EORTC QLQ-C30 consists of 30 questions assessing five aspects of participant functioning (physical, emotional, role, cognitive, and social), three symptom scales (fatigue, nausea and vomiting, pain), GHS/QoL, and six single items (dyspnea, insomnia, appetite loss, constipation, diarrhea, and financial difficulties). The symptom items were scored on a 4-point scale that ranged from "not at all" to "very much". Raw average scale scores were linearly transformed to range 0-100 with higher scores indicating higher response levels (i.e. higher symptom severity). Clinically meaningful improvement was defined as a participant with at least a 9 point scale score decrease.

The FACT-Lym subscale has 4 domains: physical well-being, social/family well-being, emotional well-being, functional well-being and lymphoma-specific subscale (LymS). The FACT-Lym LYMS consists of common lymphoma disease and/or treatment-related symptoms (e.g., pain, fever, swelling, night sweats, insomnia, itching, weight loss, fatigue, and loss of appetite). FACT-Lym LYMS contains 15 items scored from 0-4, where 0="Not at all" and 4="very much". Scale score range is from 0 to 60. Higher score indicates better quality of life. Clinically meaningful improvement was defined as a 3-point increase from baseline.

An AE was any untoward medical occurrence in a clinical investigation participant administered a pharmaceutical product, regardless of causal attribution. An AE can therefore be any unfavorable and unintended sign (including an abnormal laboratory finding), symptom, or disease temporally associated with the use of a medicinal product, whether or not considered related to the medicinal product.

An AE was any untoward medical occurrence in a clinical investigation participant administered a pharmaceutical product, regardless of causal attribution. An AE can therefore be any unfavorable and unintended sign (including an abnormal laboratory finding), symptom, or disease temporally associated with the use of a medicinal product, whether or not considered related to the medicinal product. Participants receiving polatuzumab vedotin may develop PN, including peripheral sensory and/or motor neuropathy. Symptoms included hypoesthesia, hyperesthesia, paresthesia, dysesthesia, discomfort, a burning sensation, weakness, gait disturbance, loss of balance, orthostatic hypotension, syncope, or neuropathic pain.

Dose modifications (interruptions/reductions) were based on physical examination findings, observed toxicities, and laboratory results obtained within 72 hours before study treatment administration. The determination of all dose modifications was made based on the investigator's assessment of ongoing clinical benefit with continuing study treatment. Dosing occurred only when a participant's clinical assessment and laboratory test values were acceptable.

Dose intensity = (total dose actually received divided by the expected total dose)*100. Expected total dose for dose intensity accounting for delay/reduction is based on the expected number of cycles between the first and last exposure of each drug. Percentage of dose intensity accounting for delay/reduction in dose by the participants in each arm is reported here.

OS was defined as the time from randomization to the death from any cause during the study. Participants who were not reported as having died at the time of analysis were censored at the date when they were last known to be alive. Participants who did not have post-baseline information were censored at the date of randomization. KM method was used to estimate median OS for each treatment arm.

Treatment emergent ADA-positive participants after drug administration were determined for participants exposed to polatuzumab vedotin. Participants positive for treatment emergent ADA were defined as the number of post-baseline evaluable participants determined to have treatment-induced ADA or treatment-enhanced ADA during the study period. Treatment-induced ADA = a participant with negative or missing baseline ADA result(s) and at least one positive post-baseline ADA result. Treatment-enhanced ADA = a participant with positive ADA result at baseline who has one or more post-baseline titer results that are at least 0.60 t.u. greater than the baseline titer result. Percentages are rounded off.