PGS is right. I read the criteria, the abstract was quite clear CR/PR met criteria of bone marrow and peripheral blood, not the other two of platelet count and clinical. Here is the criteria table from the article, you can see yourself:
Overall response rate was 44%. This included five (28%) patients who met the BM and peripheral blood morphologic criteria for CR (n=4) or PR (n=1) and 3 patients with clinical improvement, pending validation of response duration and resolution of drug-induced grade-1 thrombocytopenia.