Cleveland Clinic Researchers Develop Model to Help Predict Distant Metastases in Early-Stage Lung Cancer

Despite therapies achieving great local disease control, many patients with early-stage non-small cell lung cancer (NSCLC) treated with highly focused radiation therapy go on to develop distant metastases (DMs).

In response to this challenge, researchers at Cleveland Clinic have developed a model for predicting which patients are at higher risk of DMs using readily available parameters. This nomogram may help define subgroups for stratification in future clinical trials and identify patients who may benefit from adjuvant systemic therapies following lung stereotactic body radiation therapy (SBRT).

Predictive model helps identify candidates for further therapies

“Our nomogram is able to predict a probability of DMs at two years as a percentage,” says Steven Oh, MD, the lead author on the study and a resident at Cleveland Clinic. “And so we can use those numbers to stratify patients into high- or low-risk categories … and to identify those who are most likely to benefit from systemic therapy such as chemotherapy following radiotherapy.”
Results on the research behind the nomogram will be presented at the 2015 annual meeting of the America Society for Radiology Oncology (ASTRO) in San Antonio, Texas. The model was validated internally and external validation is warranted, the abstract concludes.

Cleveland Clinic researchers examined their institutional registry of patients treated with lung SBRT between 2003 and 2014 and identified 729 patients with early-stage NSCLC eligible for analysis. Of those patients, 157 were found to have developed distant metastases at a median time of 10.3 (range 0.2-68.4) months.

The standard of treatment for early-stage NSCLC is surgical resection of the lobe where the tumor is located. However, many patients are not surgical candidates because of medical comorbidities. So, physicians have turned to SBRT, using high doses of focused radiation to treat patients who otherwise would not be able to receive any treatment.

“Depending on the study, somewhere between 25 to 33 percent — essentially a fourth or more of the patients — are medically inoperable. So for this portion of patients the role of SBRT has increased,” says Dr. Oh.

Successful local control with SBRT does not stop cancer’s spread

“Our institution has a large experience in treating patients with this technique, and the good news is that local control is excellent with this modality — it’s over 95 percent with regards to local control of this disease,” Dr. Oh says. “However, many patients after they have this procedure develop DMs and their cancer spreads to … either another part of the lung, their other lung, or to their brain, liver or the adrenal glands. So the major problem … is that many patients will develop metastatic disease despite the local control.”

The same is true for many patients who are able to undergo surgery and don’t have other medical comorbidities, Dr. Oh says.

“They also have a rate of DMs somewhere between 15 to 17 percent,” he says. “In our study, we looked at 729 of our patients treated here at Cleveland Clinic where we use SBRT, and we found that about 22 percent developed distant metastases despite being treated. And so the question is, can we predict who will develop distant metastases?”

The ability to predict which patients are likely to develop DMs is important, because chemotherapy is effective in treating such patients. “However, we currently do not give chemotherapy after early-stage lung cancer treated with SBRT,” Dr. Oh says. “And the main problem is that these patients who would be candidates for chemotherapy because of their DMs are often too sick to receive chemotherapy, they have too many underlying medical conditions.”

In working with the model, however, researchers found that the patients more likely to develop distant metastases tended to be among the healthier patients who may be able to tolerate chemotherapy.

“What we were able to show with our nomogram is that the patients who were at the highest likelihood of developing DMs are in fact the healthier patients who have less comorbid conditions, and this is likely because they are not in fact dying from their comorbid conditions, so they have longer time and a higher chance of developing the distant metastases,” Dr. Oh says.

Nomogram works with patient information that is readily available

The model doesn’t require any additional testing that the patient would not typically have. Among the variables included for each patient are:
• Age
• Body mass index
• Measurement of comorbid conditions (taken from medical history)
• Tumor size and activity measured by CAT and PET scans

“All of these variables can be extracted just from the patient’s medical history … it just requires essentially plugging these numbers into a model,” says Dr. Oh.
The project is significant in that it shows that people who are otherwise healthy but have unfavorable tumor characteristics, such as a large tumor or a very active tumor, are the patients at higher risk of developing distant failure, Dr. Oh says.

“Our hope is that we can better stratify patients for future clinical trials, and possibly, in the future, design a trial where patients who are otherwise healthy and are candidates for chemotherapy can be identified,” he says.