1. Added optional biomarker analysis as additional efficacy endpoint, including SPARC testing, of tumor tissue and blood has in order to further study the correlation between expression of molecular biomarkers and clinical outcome. 2. In response to the FDA Clinical Pharmacology Reviewer’s recommendation, added sparse PK sampling and sampling rationale.

1. References to dacarbazine brand names have been removed throughout the document. 2. Language added to clarify that participation in PK sampling is optional. 3. Added clarification that patients who stopped treatment prior to developing disease progression were to be followed without further treatment until progressive disease was documented or until the treating physician felt that additional treatment was required (in which case documentation was to be made in the Case Report Form as to the compelling reason(s) for starting a new treatment in the absence of progressive disease). 4. Added procedures for collection and preparation of specimens for biomarker analyses. 5. Updated storage and preparation information for dacarbazine. 6. The Every-8-Weeks assessment of CBC, differential and platelet count has been removed from both arms as it was erroneous. 7. Added clarification that PK was an optional procedure. 8. The list of tests to be included in the clinical chemistry panel was detailed. 9. Sponsor Signatories page has been updated.

1.Change of Medical Monitor 2. Clarified that pharmacokinetic evaluations are for the ABI-007 arm only. 3. Added clarification that patients should have cutaneous malignant melanoma for entry into the study. 4. Updated language to reflect expression of SPARC in metastatic melanoma and its association with a poor prognosis. 5. Added details regarding premedication for dacarbazine arm (Arm B). 6. Deleted paragraph regarding patients who received chemotherapy or an investigational drug 3 weeks prior to first dose since prior cytotoxic chemotherapy and prior adjuvant cytotoxic chemotherapy were not permitted for patients in this study. 7. CT scan or MRI of head assessment added for consistency. 8. Added outline of planned interim safety review. 9. Added a section to outline the establishment of a Data Monitoring Committee

1. Changed time window for dacarbazine infusion from 15 – 60 minutes to 30 – 60 minutes. 2. The following changes were made to inclusion criteria: - Inclusion Criterion 2 was updated to clarify the timeframe by which sites should allow a patient to enter the study upon completing prior treatment. - Inclusion Criterion 3 was updated to clarify the guidelines regarding contraception during and after treatment in the protocol, and to be consistent with the Summary of Product Characteristics and Prescribing Information. - Inclusion Criterion 4 was updated to clarify the length of time and types of previous malignancies that are and are not acceptable for entry into the study. - Inclusion Criterion 5 was modified with language to reinforce that any patients who enrolled into the trial were healthy enough to complete the study. The following changes were made to exclusion criteria: - Exclusion 8 was added to ensure that patients with hypersensitivity reactions to any of the investigational products or comparators were not enrolled. 3. Text for blood collection for biomarkers was modified to ensure that sites were allowed to minimize any extra stress put on the patient by having multiple blood draws in a compressed time frame. 4. Clarification was added regarding participation in other investigational trials while participating in the current study, and also regarding radiotherapy with reference to Treatment for Brain Metastases section. 5. Language was updated to allow sites to use their standard practice for study drug administration while ensuring that any major changes in body weight were accounted for in the BSA calculations for dosing the patient. 6. Added BSA calculation at Baseline. 7. Added guidance for abnormal lymph nodes identified as target lesions, for target lesions documented via digital photography and for the assessment of progressive disease for pleural fluid, ascites, pericardial effusions, and other fluid collections.

1. Changes were made throughout the protocol to reflect the acquisition of Abraxis BioScience by Celgene Corporation. 2. Added text to clarify the acceptable window for ABI-007 infusion time (30 – 40 minutes). 3. Added language to clarify that because pharmacokinetic information was not being collected for the dacarbazine treatment arm in this study, the infusion time window for dacarbazine could be determined by guidelines followed at the clinical site. 4. Removed text specific to United States and Canada from the PK sections to allow study sites from additional regions to participate. 5. Added clarification that PK sampling participation was to be determined by the patient. 6. Inclusion criteria were modified to allow enrollment of patients with other malignancies if they were cured by surgery with radiotherapy. 7. Text modified to clarify that the biomarker assays were part of a sub-study involving a subset of patients and that the assays performed were for exploratory purposes and were not validated. 8. Added clarification that follow-up of non-serious AEs will not continue if patient initiates another anti-cancer therapy.

1. Updated the IND number on the protocol to IND 115025. IND 115025 administratively split the malignant melanoma indication from the existing IND 055974 for ABI-007 at the request of FDA as a result of the reorganization at the FDA Office of Oncology Drug Products in September 2011. 2. Updated Contact Information to reflect the acquisition of Abraxis BioScience by Celgene Corporation. 3. Added a final patient contact for follow-up OS status. 4. Clarified that lack of efficacy (disease progression) is not considered an AE or SAE. 5. An administrative change to update the safety reporting contact information and the methods for reporting to Celgene Drug Safety instead of Abraxis. 6. Updated the Study Design, Randomization, Sample Size, Power, and Interim Safety Review section (Section 8.1) to reflect the change in timing (and rationale) for the final analysis of the primary endpoint of independently assessed PFS as a result of independent statistical review. 7. Independent review of pooled PFS events revealed that the final event total would be lower than originally planned 379. Thus the final analysis of PFS was performed one year after the last patient was randomized (data cutoff date of 30 Jun 2012). Based on a projected 320 total PFS events, the power to reject the null hypothesis that the ABI-007/dacarbazine HR for PFS is equal to 1.0 will be approximately 72% for HR = 0.750 utilizing the remaining two-sided alpha of 0.049 for the final analysis instead of at least 80% (with a two-sided type I error of 0.049). 

1. An administrative change to reflect approval of study drug for the indication of non-small cell lung cancer patients in the US since last protocol revision. 2. Updated OS follow-up to continue to collect survival status beyond 24 months from treatment discontinuation, until death or study termination in all patients. 3. Safety reporting procedures have been broadened to reflect the extended overall survival follow-up period. 4. Administrative change to update signature page with personnel at Celgene Corporation.