This is a lengthy review published in JCO and is briefly described below. Cervical cancer remains the fourth most common cancer in women, with 527,000 new cases in 2012, and the 1999 NCI recommendation for treatment of focal advanced cervical cancer (IB2-IVA) should include platinum-based chemotherapy in addition to radiation therapy, which reduces the risk of death by 40%. Even so, the recurrence rate reaches 20% and the 3-year overall survival rate fluctuates between 67% and 74% (see Table 1). How to prevent recurrence remains the focus of research. Currently, except for the study by Due?as-González et al, which found that radiotherapy/cisplatin + gemcitabine followed by two cycles of cisplatin/gemcitabine treatment was superior to radiotherapy (Duenas-Gonzalez A, Zarba JJ, Patel F, et al. J Clin Oncol. 2011;29(13):1678-85.), no no other strategy has been found to outperform radiotherapy in terms of remission rates and survival. In early studies, 50% of patients treated with radiotherapy never relapsed. For these patients, additional treatment may only increase the harms. The addition of cisplatin chemotherapy reduced the overall recurrence rate by another 15 percent. How to ensure that the remaining 35% of patients receive optimal treatment and avoid chemotherapy toxicity in the remaining 65% of the cancer population is extremely important for maximizing and optimizing OS and quality of life. Although prognostic factors for recurrence have been well studied, how these factors apply to widely administered radiotherapy remains elusive, and Rose et al. designed a line graph of cervical cancer prognosis based on the GOG phase III study (see yesterday’s report, Rose PG, Java J, Whitney CW, et al. J Clin Oncol. 2015 .), promising better individualized estimates of patient risk and targeted, more intensive treatment for high-risk patients. Even so, decisions still need to be made in the context of the individual patient, including patient willingness to treat, treatment goals, comorbidities, and other factors outside of predictive models that are equally important for patients to receive optimal treatment. The integration of this information is key. In addition, the research base of this predictive model needs to be considered, such as the duration, dose, and method of radiotherapy in the study. For developed regions where 3D conformal radiotherapy, intensity-modulated radiotherapy and volume dose calculation have been achieved, this column line graph needs to be applied with caution. In contrast to early-stage cervical cancer, stage IVB/recurrent cervical cancer is often incurable.GOG 169 suggested that the addition of paclitaxel to cisplatin improved PFS and remission rates, GOG 204, GOG 240, and JCOG 0505 (see previous day’s report, Kitagawa R, Katsumata N, Shibata T, et al. J Clin Oncol. 2015.) explored treatment strategies for stage IVB/recurrent cervical cancer.GOG compared cisplatin/paclitaxel and cisplatin/other drug chemotherapy regimens to determine the therapeutic status of cisplatin/paclitaxel in this patient population.GOG 240 explored the effect of the addition of bevacizumab and the effect of paclitaxel/topotecan. The addition of bevacizumab significantly improved OS (17.0 vs. 13.3 months). In Katagawa’s study, paclitaxel/carboplatin was not inferior to paclitaxel/cisplatin, except in patients with no prior platinum application. Therefore, three chemotherapy regimens are currently available for patients with stage IVB/recurrent cervical cancer: paclitaxel/carboplatin, paclitaxel/cisplatin, and paclitaxel/cisplatin/bevacizumab. In order to achieve optimal treatment for each patient, remission of patient symptoms and quality of life, and prolonging survival as much as possible should be important goals. Table 2 compares the three aforementioned studies. Among them, GOG 240 had the most stringent inclusion criteria, such as the requirement of good renal function (creatinine clearance > 60 ml/min, tightly controlled hypertension, etc.). Expected chemotherapy toxicity is also an important consideration in the choice of treatment for patients with underlying comorbidities and previously treated patients. Overall, both paclitaxel/carboplatin and paclitaxel/cisplatin were easily accepted, especially with the change of the paclitaxel intravenous infusion time to 3 hours and the appropriate antiemetic treatment for cisplatin-treated patients. When considering expected toxicity, it should be considered that the majority (83%) of patients had received prior radiotherapy, which significantly increased the hematologic toxicity and specific bevacizumab-related toxicity of the treatment. the incidence of febrile neutropenia in the paclitaxel/cisplatin-treated groups in GOG 204, GOG 240, and JCOG 0505 was 12.9%, 5%, and 16%, respectively. The lower incidence of febrile neutropenia associated with paclitaxel/carboplatin in JCOG 0505 (7.1%) may be related to the modification of the initial dose of carboplatin (adjusted to AUC 5). Treatment-related fistula formation is associated with previous radiotherapy and disease recurrence sites and should be discussed with patients to account for it. This is because 54% and 38% of patients in GOG 240 and JCOG 0505, respectively, had a recurrence site in the pelvis or in the radiotherapy field. In GOG 240, the incidence of gastrointestinal fistula was 8.3% vs. 0.9% with and without bevacizumab, respectively. All fistulas occurred in patients with prior radiation therapy. There are no predictive line graphs for the effect of the number of chemotherapy drug classes (two or three) on treatment outcome, and prognostic criteria need to be determined based on previous studies. In the Moore study (Moore DH, Tian C, Monk BJ, et al. Gynecol Oncol. 2010;116(1):44-49.), a retrospective assessment of prognostic factors from patients with stage IVB, recurrent, or progressive cervical cancer found the following factors to reduce tumor response to chemotherapy: black race, performance status ≥0, prior cisplatin history of treatment, disease confined to the pelvis, and recurrence within 1 year. Accordingly, patients were divided into three groups: low-risk (0-1 high-risk factors), intermediate-risk (2-3 high-risk factors), and high-risk (4-5 high-risk factors). High-risk patients were 13% effective for platinum-based chemotherapy, while low-risk patients were predicted to be 51% effective. These criteria were tested for Tewari in GOG 240, confirming the predictive value of these factors. In GOG 240, high-risk patients benefited the most from treatment with bevacizumab (OS: 6.3 vs. 12.1 months), while the low-risk group did not benefit significantly (OS: 23.0 vs. 21.8 months). The estimated HR for death with bevacizumab was 0.96,0.67 and 0.54 in the low-risk, median and high-risk patients, respectively, compared to the no-bevacizumab population. Health resource limitations are also factors to consider, with the addition of bevacizumab costing up to $155,000/quality-adjusted life-year in GOG 240. Although low-risk patients could also benefit from the addition of bevacizumab (52% vs. 63% efficiency), the cost in the low-risk group would be well above the general average, considering that bevacizumab is administered until disease progression.