This randomized, Phase III, multicenter, open-label study designed to evaluate the efficacy of Atezolizumab in combination with carboplatin, paclitaxel, bevacizumab compared with treatment with pemetrexed, cisplatin in approximately 228 TKI(tyrosine kinase inhibitor) pre treated patients with Stage IV non squamous non small cell lung cancer with activating EGFR mutation or ALK translocation.

Inclusion Criteria

• Male or female, 18 years of age or older
• Eastern Cooperative Oncology Group(ECOG) performance status of 0 or 1
• Histologically or cytologically confirmed, Stage IV non-squamous NSCLC (per the Union Internationale contre le Cancer/American Joint Committee on Cancer staging system, 8th edition).
• Patients with tumors of mixed histology (i.e., squamous and non-squamous) are eligible if the major histological component appears to be non-squamous.
• As the stage IV non-squamous NSCLC treatment, no prior cytotoxic treatment is allowed and the patients treated with below tyrosine kinase inhibitor prior to the enrollment
• Patients with a sensitizing mutation in the epidermal growth factor receptor (EGFR) gene must have experienced disease progression (during or after treatment) or intolerance to treatment with one or more EGFR TKIs, such as erlotinib, gefitinib, osimertinib or another EGFR tyrosine kinase inhibitor (TKI) appropriate for the treatment of EGFR mutant NSCLC.

If the patients have identified T790M mutation after 1st or 2nd generation EGFR TKI failure, the patient must be treated with the second line 3rd generation EGFR TKI treatment, such as osimertinib, before the study participation.

Patients with an anaplastic lymphoma kinase (ALK) fusion oncogene must have experienced disease progression (during or after treatment) or intolerance to treatment with one or more ALK inhibitors (i.e., crizotinib) appropriate for the treatment of NSCLC in patients having an ALK fusion oncogene.

Patients with unknown EGFR and/or ALK status require test results at screening. ALK and/or EGFR may be assessed locally or at a main investigator laboratory.

Inoperable stage III non-squamous NSCLC treatment, no prior cytotoxic treatment is allowed and the patients treated with below tyrosine kinase inhibitor prior to the enrollment

• Patients who have received prior neo-adjuvant, adjuvant chemotherapy, radiotherapy, or chemoradiotherapy with curative intent for non-metastatic disease must have experienced a treatment free interval of at least 6 months from randomization since the last chemotherapy, radiotherapy, or chemoradiotherapy.
• Patients with a history of treated asymptomatic Central Nervous System(CNS) metastases are eligible, provided they meet all of the following criteria:

No ongoing requirement for corticosteroids as therapy for Central Nervous System(CNS) disease No stereotactic radiation within 7 days or whole-brain radiation within 14 days prior to randomization No evidence of interim progression between the completion of Central Nervous System(CNS)-directed therapy and the screening radiographic study

• Patients with new asymptomatic Central Nervous System(CNS) metastases detected at the screening scan may be eligible without the need for an additional radiation therapy and/or surgery for Central Nervous System(CNS) metastases by investigator's decision.
• Patients who are available to provide tissue from previously obtained archival tumor tissue or tissue obtained from a biopsy at screening for the PD-L1 test.

( in case, if patient cannot provides any slides, the enrollment can still be considered with individual discussion with the principal investigator)

- Biopsy and blood sample requirement are as follows for the Whole exome and transcriptome sequencing: A representative formalin-fixed paraffin-embedded (FFPE) tumor specimen in paraffin block (preferred) or 10 or more unstained, freshly cut, serial sections on slides from an FFPE tumor specimen is required for participation in this study. If fewer than 10 slides are available, patient can be still considered for the enrollment based on the discussion and the decision with the principal investigator. This specimen must be accompanied by the associated pathology report.

Fine-needle aspiration (defined as samples that do not preserve tissue architecture and yield cell suspension and/or cell smears), brushing, cell pellet specimens (e.g., from pleural effusion, and lavage samples) are not acceptable.

Tumor tissue from bone metastases that is subject to decalcification is not acceptable.

Endobronchial ultrasonographic biopsy samples are not preferred but acceptable. For core needle biopsy specimens, preferably at least three cores embedded in a single paraffin block, should be submitted for evaluation.

30ml of peripheral blood sample acquired during the screening period

- Measurable disease, as defined by RECIST v1.1 Previously irradiated lesions can only be considered as measurable disease if disease progression has been unequivocally documented at that site since radiation and the previously irradiated lesion is not the only site of disease.

Adequate hematologic and end organ function, defined by the following laboratory results obtained within 14 days prior to randomization:

ANC ≥ 1500 cells/µL without granulocyte colony stimulating factor support Lymphocyte count ≥500/µL Platelet count ≥ 100,000/µL without transfusion Hemoglobin ≥ 9.0 g/dL Patients may be transfused to meet this criterion. INR or aPTT ≤ 1.5 x upper limit of normal (ULN) This applies only to patients who are not receiving therapeutic anticoagulation; patients receiving therapeutic anticoagulation should be on a stable dose.

AST, ALT, and alkaline phosphatase ≤ 2.5 x ULN, with the following exceptions:

Patients with documented liver metastases: AST and/or ALT ≤ 5 x ULN Patients with documented liver or bone metastases: alkaline phosphatase ≤ 5 x ULN.

Serum bilirubin ≤ 1.25 x ULN Patients with known Gilbert disease who have serum bilirubin level ≤ 3 x ULN may be enrolled.

Serum creatinine ≤ 1.5 x ULN

• For female patients of childbearing potential, agreement (by patient and/or partner) to use a highly effective form(s) of contraception that results in a low failure rate (< 1% per year) when used consistently and correctly, and to continue its use for 5 months after the last dose of atezolizumab and/or 6 months after the last dose of bevacizumab or paclitaxel, whichever is later). Such methods include: combined (estrogen and progestogen containing) hormonal contraception, progestogen-only hormonal contraception associated with inhibition of ovulation together with another additional barrier method always containing a spermicide, intrauterine device (IUD): intrauterine hormone-releasing system (IUS), bilateral tubal occlusion, vasectomized partner (on the understanding that this is the only one partner during the whole study duration), and sexual abstinence.
• For male patients with female partners of childbearing potential, agreement (by patient and/or partner) to use a highly effective form(s) of contraception that results in a low failure rate [< 1% per year] when used consistently and correctly, and to continue its use for 6 months after the last dose of bevacizumab, carboplatin, or paclitaxel. Male patients should not donate sperm during this study and for at least 6 months after the last dose of bevacizumab, carboplatin, or paclitaxel.
• Oral contraception should always be combined with an additional contraceptive method because of a potential interaction with the study drug. The same rules are valid for male patients involved in this clinical study if they have a partner of childbirth potential. Male patients must always use a condom.
• Women who are not postmenopausal ( 12 months of non therapy-induced amenorrhea) or surgically sterile must have a negative serum pregnancy test result within 14 days prior to initiation of study drug

Exclusion Criteria

Patients who meet any of the criteria below will be excluded from study entry. Cancer-Specific Exclusions Active or untreated symptomatic Central Nervous System(CNS) metastases as determined by CT or magnetic resonance imaging (MRI) evaluation during screening and prior radiographic assessments Spinal cord compression not definitively treated with surgery and/or radiation or previously diagnosed and treated spinal cord compression without evidence that disease has been clinically stable for > 2 weeks prior to randomization Leptomeningeal disease Uncontrolled tumor-related pain Patients requiring pain medication must be on a stable regimen at study entry. Symptomatic lesions amenable to palliative radiotherapy (e.g., bone metastases or metastases causing nerve impingement) should be treated prior to randomization. Patients should be recovered from the effects of radiation. There is no required minimum recovery period.

Asymptomatic metastatic lesions whose further growth would likely cause functional deficits or intractable pain (e.g., epidural metastasis that is not currently associated with spinal cord compression) should be considered for locoregional therapy, if appropriate, prior to randomization.

Uncontrolled pleural effusion, pericardial effusion, or ascites requiring recurrent drainage procedures (once monthly or more frequently) Patients with indwelling catheters (e.g., PleurX®) are allowed. Uncontrolled or symptomatic hypercalcemia (> 1.5 mmol/L ionized calcium or Ca > 12 mg/dL) Malignancies other than NSCLC within 5 years prior to randomization, with the exception of those with a negligible risk of metastasis or death (e.g., expected 5 year OS > 90%) treated with expected curative outcome (such as adequately treated carcinoma in situ of the cervix, basal or squamous-cell skin cancer, localized prostate cancer treated surgically with curative intent, ductal carcinoma in situ treated surgically with curative intent)