The effectiveness of As2O3 treatment was studied in 3 carcinoma cell lines, LoVo, OVCAR-3, and PA-1, and in cancer cells obtained from ascites fluids of 8 patients with carcinomatosis peritonei. LoVo, OVCAR-3, and PA-1 cell lines, and cancer cells from the patients were cultured in As2O3 gradient media; As2O3 sensitivity was evaluated by trypan blue dye exclusion and by morphologic examination after Wright staining. PA-1 was the most sensitive cell line to As2O3; OVCAR-3 and LoVo were resistant to As2O3. Cancer cells from 2 of 8 patients were sensitive to As2O3. The in vivo tumoricidal effect of As2O3 (100 microg/day, i.p.) was studied in 30 BALB/c nude mice following i.p. implantation of PA-1 tumor cells. The 17 As2O3-injected mice died of extensive intratumoral hemorrhage, necrosis, and hemorrhagic ascites within 48 hr after initial treatment. In 10 As2O3-untreated tumor-bearing control mice, only focal intratumoral hemorrhage and necrosis were noted. In summary, solid tumor cell lines and cancer cells from patients showed various As2O3 sensitivities in vitro, and As2O3 had a marked tumoricidal effect on PA-1 cells in vivo. These results suggest that As2O3 treatment might possibly be beneficial in patients with carcinomatosis peritonei who are resistant to conventional therapy and whose tumors show in vitro sensitivity to As2O3. However, to minimize the life-threatening tumor lysis effect, it would be better to administer As2O3 after removal of the peritoneal tumor masses.