Regorafenib is approved in the treatment for metastatic colorectal cancer patients who have been progressed after standard therapies, however, there has not been a predictive biomarker. The investigators designed this study to investigate whether [18F]FLT-PET might paly a role as a predictive imaging biomarker of treatment responses to regorafenib.

Recent advances have been made in the treatments for the patients with metastatic colorectal cancer (mCRC) owing to the introductions of targeted agents, which included bevacizumab, cetuximab, panitumumab, and aflibercept. And in addition, regorafenib, a newer tyrosine kinase inhibitor (TKI), has been approved in the treatment for the mCRC patients.

Regorafenib (BAY 73-4506) is an orally available multikinase inhibitor with activity against multiple targets, including tumor angiogenesis (VEGFR-1, -2, -3 and TIE-2), oncogenesis (KIT, RET, RAF-1, BRAF, and BRAFV600E), and tumor microenvironment (PDGF and FGFR). Regorafenib has shown antitumor activities in multiple solid tumors, and demonstrated significant efficacy outcomes in patients with advanced gastrointestinal stromal tumors and colorectal cancers.

The CORRECT study, which compared regorafenib vs placebo in mCRC patients who have been treated with all standard treatment, showed survival improvements with statistical significances; median OS 6.4 vs 5.0 months, HR 0.77, p=0.0052; median PFS 1.9 vs 1.7 months, HR 0.49, p<0.000001. Above these results, regorafenib monotherapy has been recently approved for the treatment of mCRC patients who have been refractory to all of standard therapies.

However, there are still only a few biomarkers which have been established as predictive of treatment responses in the fields of treatments for mCRC patients; KRAS or BRAF mutations for the lack of responses to anti-EGFR agents, cetuximab or panitumumab. There still has not been any biomarker which would be predictive of treatment responses to bevacizumab, aflibercept or regorafenib. The difficulties in search for biomarkers for these agents might come from the facts as following; either bevacizumab or aflibercept does not act directly against tumor itself and should be combined with cytotoxic agents to show efficacy; regorafenib is a multikinase inhibitor which has too many potential targets.

Above these reasons, imaging modalities can be fascinating and alternative candidates for predictive biomarkers of treatment responses. Conventional anatomic imaging studies such as computed tomography (CT) scans can hardly predict the treatment responses earlier, and the RECIST using CT scans, which is widely used for measurement of treatment responses, might have several limitations for measurement of efficacy from targeted agents such as cystic necrosis without tumor shrinkage. In the CORRECT study, overall response rate by RECIST was only 1%, although the rates for disease stabilization was up to 40%, which might be a good example for the limitations of the RECIST using conventional anatomical imaging studies for the response evaluation of regorafenib.

Among imaging studies, PET scans are useful tools for the noninvasive measurement of functional changes after treatment with targeted agents, and [18F]FLT-PET is potentially useful tool for earlier prediction of treatment responses because it can detect earlier changes of cellular proliferation using [18F]FLT (fluorothymidine), a radiotraceable substitute for thymidine which is essential for DNA synthesis. Several studies have been reported that [18F]FLT-PET may allow an early assessment of the response to chemotherapy including targeted agents. There also has been a report that [18F]FLT-PET could predict treatment responses of BRAF inhibitors in the colorectal cancer xenograft model; regorafenib also has an inhibitory effect on BRAF.

Therefore, the investigators have planned this study with hypothesis that [18F]FLT-PET could be useful for identifying a subgroup of mCRC patients with clinical responsiveness to regorafenib.

Inclusion Criteria:

1. Histologically or cytologically confirmed adenocarcinoma of the colon or the rectum.
2. Progressed after 3 active cytotoxic chemotherapy including fluoropyrimidines, oxaliplatin and irinotecan during or within 6 months of their administrations with or without targeted agents (bevacizumab or cetuximab).
3. Extrahepatic measurable lesion(s) by RECIST 1.1.
4. Unresectable metastatic disease.
5. Age over 20 years old.
6. Have a life expectancy of at least 3 months.
7. ECOG performance status of 1 or lower.
8. Adequate organ functions.
9. Be willing and able to comply with the protocol for the duration of the study.
10. Give written informed consent prior to study-specific screening procedures, with the understanding that the patient has the right to withdraw the study at any time, without prejudice.
11. Women of childbearing potential and men must agree to use adequate contraception since signing of the IC form until at least 8 weeks after the last study drug administration.

Exclusion Criteria:

1. Prior treatment of regorafenib.
2. Liver-limited metastasis.
3. Inability to perform [18F]FLT and [18F]FDG-PET imaging studies due to physical inability or claustrophobia.
4. Concurrent or previous history of another primary cancer within 3 years prior to randomisation except for curatively treated cervical cancer in situ, non-melanomatous skin cancer, superficial bladder cancer (pTis and pT1) and curatively treated thyroid cancer of any stage. Concurrent, histologically confirmed, unresected thyroid cancer without distant metastasis could be allowed with the agreement of the chief principal investigator.
5. Uncontrolled CNS metastases.
6. Prior radiation therapy would be permitted, but non-radiated evaluable lesions should be present at study entry.
7. Uncontrolled hypertension (>150/90 mmHg) despite of optimal management; anti-hypertensive drugs for BP lowering before study entry would be permitted.
8. Congestive heart failure ≥ New York Heart Association (NYHA) class 2.
9. Unstable angina, new-onset angina within 3 months, or history of myocardial infarction within 6 months before the study entry.
10. Arterial or venous thromboembolism within 6 months.
11. Serious concurrent infections or non-malignant illness.
12. Liver cirrhosis ≥ Child-Pugh class B.
13. Active hepatitis B or C, or chronic hepatitis B or C requiring treatment with antiviral therapy.
14. Peripheral neuropathy of grade ≥ 2.
15. Major surgery or significant traumatic injury within 28 days prior to study treatment.
16. Non-healing wound, ulcer, or bone fracture.
17. Current evidence of significant gastrointestinal bleeding or (impending) obstruction.
18. Any hemorrhage or bleeding event of grade ≥ 3 within 4 weeks prior to the start of study medication.
19. Proteinuria ≥ 3+ in the routine urinalysis; in this case, the total protein in the 24-hour urine collection should be measured, and the accrual is permitted if total protein < 3.5 g/day.
20. Pregnant of breast-feeding subjects. Women of child-bearing potential must have pregnancy test within 7 days and a negative result must be documented before start of study treatment.
21. Substance abuse, medical, psychological or social conditions that may interfere with the subject's participation in the study or evaluation of the study results.
22. Use of strong CYP3A4 inducers or inhibitors which are known to decrease the metabolism of regorafenib (ketoconazole, rifampin, phenytoin, carbamazepine, phenobarbital).
23. Known hypersensitivity to the study drug or any of its excipients.

Between July 2014 and June 2016, we enrolled 68 patients with mCRC who showed progression after all standard therapies, and initiated Regorafenib administration (160 mg/day) at Asan Medical Center in Seoul, Republic of Korea.