Comprehensive genomic profiling (CGP) for cancer has been introduced into health insurance practice in Japan since June 2019. The panel tests are indicated for those with advanced solid tumors who are in good condition to receive cancer treatment, typically as in a clinical trial, and who have completed or are likely to complete standard treatment soon. At government-designated core base hospitals or base hospitals, an intra-institutional molecular tumor board, consisting of a multidisciplinary medical team called an expert panel, annotates the CGP test results to propose appropriately targeted treatments for each patient. However, the treatment options are suggested in only approximately 10% of patients, implying that CGP testing is of no use to most patients. Furthermore, the options proposed are usually limited to treatments in clinical trials or that require the use of off-label drugs that are not covered by health insurance. Therefore, even the few privileged patients who receive treatment are faced with difficulties. Given this situation, it is clear that the most critical issue in current cancer genomic medicine is an exit strategy: how to link more patients to appropriate treatments. Realistically, CGP testing has made targeted therapy under clinical trials, with an unconfirmed efficacy and safety, a tempting treatment option for patients. Therefore, to benefit patients, medical oncologists must reasonably deal with such conflicts, distinguishing between evidence-based and experimental treatments.