Increasingly, medical oncologists work with multidisciplinary teams to diagnose and manage lung cancer cases. Interventional pulmonologists on these teams perform diagnostic workups, obtain tissue biopsies, and play a role in the timely diagnosis and staging of lung cancer (Ann Am Thorac Soc 2015;12(5):742-751). The number of molecular tests that can impact patient treatment is increasing, while biopsies are becoming less invasive, resulting in less tissue to work with.
Pulmonologists are working with pathologists to design tissue-use algorithms and verify that biopsies will be sufficient for molecular testing. Thus, pulmonologists are in a unique position to order prognostic and predictive biomarkers for all stages of non-small cell lung cancer (NSCLC) and use this information with their medical oncologists and other members of their multidisciplinary team (Am J Respir Crit Care Med 2013;188(4):503-507).
The gold standard for identification of tumor drivers is molecular analysis of tissue, taken during a biopsy procedure. However, these analyses can often take 1-2 weeks or longer depending on how samples are retrieved and what testing is being conducted. Additionally, samples that are inadequate for molecular analysis may require a repeat biopsy, resulting in increased time to treatment and a number of procedures for the patient.
In contrast, blood-based liquid biopsies that analyze cell-free DNA and the circulating proteome are easily obtainable with a simple blood draw and can be done either independent of, or concurrent with, a tissue biopsy. Actionable molecular information obtained from blood is highly concordant with tissue, with sensitivities ranging from 85 percent to 96 percent (J Mol Diagn 2017;19(3):404-416). Turnaround times are much shorter than for tissue biopsies with results available in as little as 2-3 days. Liquid biopsies can be ordered by interventional pulmonologists or thoracic surgeons at the initial diagnostic workup, and results can be presented at the initial multidisciplinary meeting for a patient. This information can be reviewed and used by the treating medical oncologist and team prior to initiation of treatment to design an appropriate and personalized therapeutic regimen.
Utilization of prognostic and predictive markers, such as driver mutations that predict targeted therapy response or proteomic profiling, are important in developing personalized monitoring schedules and therapeutic regimens, as well as facilitating prognostic conversations with patients. Early identification of actionable tumor drivers, such as EGFR mutations or ALK and ROS1 translocations, is critical to getting patients on the proper targeted therapy or clinical trial in the front-line setting. If treatment begins without these results, patients with driver mutations are at risk for beginning treatment with a non-targeted, systemic therapy, which data demonstrate decreases overall survival (JAMA 2014; 311(19):1998-2006).
Case Study 1
Liquid biopsy identifies predictive biomarkers and reduces time to treatment
In April 2017, a 77-year-old female with no smoking or lung cancer history presented with a dry cough and no other symptoms. High-resolution CT scan showed abnormalities and a subsequent PET scan showed multiple hypermetabolic pleural-based nodules on the right lung. A liver biopsy and a blood draw for liquid biopsy were completed on the same day for both proteomic and genomic, predictive and prognostic markers. The patient was diagnosed with metastatic adenocarcinoma with clear cell features, ECOG PS 0, and the liquid biopsy revealed an EGFR-activating mutation.
Based on this information, the patient received her first dose of erlotinib at her initial medical oncology visit, 8 days after first abnormal CT scan. Tissue biopsies later confirmed the presence of an EGFR-activating mutation; however, those results were not available for 24 days after initial diagnostic workup. In this case, sole reliance on the tissue biopsy would have created a 3-week delay in identification of activating mutations. During this time, either treatment would be delayed while waiting for molecular results, or the patient would have started on front-line chemotherapy and been switched to targeted therapy when mutation results were received.
The patient is currently continuing on erlotinib and scans show regression of disease.
Liquid biopsies, as opposed to tissue biopsies, allow for easy serial monitoring for changes in predictive mutations in response to targeted therapy. Disappearance of predictive mutations from the blood after initiation of targeted therapy can be indicative of favorable therapeutic response (Clin Cancer Res 2015;21(14):3196-3203); likewise, a reappearance of the mutation may indicate that the patient is no longer responding. Serial monitoring can also provide early detection of resistance mutations that may indicate a need to switch therapy.
To obtain the most comprehensive picture of a patient's disease, prognostic information should also be obtained in addition to predictive mutation drivers. Multi-variate, protein-based liquid biopsies that can classify patients into "good" or "poor" prognostic categories are available (Br J Cancer 2017;116(1):36-43, Curr Med Res Opin 2017;33(6):1091-1097).
When used in conjunction with other clinical factors, such as smoking status, performance status, age, and comorbidities, these tests can help identify "poor" prognostic patients that will underperform on standard therapeutic regimens. These patients may be candidates for more aggressive treatment regimens, such as immunotherapy and chemotherapy combinations, more frequent monitoring to catch progression early, or, in later lines of therapy, palliative care. The multidisciplinary team can review all factors and decide if a patient with few co-morbidities and a good performance status, but poor proteomic prognosis, is a candidate for a clinical trial. Or perhaps an aggressive systemic therapy regimen? For patients who do not want to pursue therapy, can a good proteomic prognosis be used to encourage patients toward treatment because it could improve their outcomes? Understanding all prognostic factors for a patient will help the team build a personalized treatment plan to help care for the individual patient (Curr Med Res Opin 2017;33(6):1091-1097).
Historically, prognostic and predictive marker results were mostly used in advanced stages of NSCLC; however, as the treatment landscape evolves, they are quickly becoming relevant in earlier stage disease. Various therapeutic options, either targeted or not, are available in the neoadjuvant and adjuvant settings, and the data on the interplay between mutations and response to radiation therapy are evolving. Prognostic information gleaned from liquid biopsies can identify patients with more aggressive disease and inform surgical and adjuvant treatment planning.
Case Study 2
Early stage patient with multiple poor prognostic factors detected by liquid biopsy
A 65-year-old female with smoking history, chronic obstructive pulmonary disease, and cerebrovascular accident presented in June 2016 with significant weight loss. PET scans revealed a 2.5 cm lower left lobe (LLL) nodule and a 6 mm right upper lobe (RUL) nodule. A subsequent MRI was negative and no extrathoracic disease was identified. EBUS showed negative lymph nodes, LLL nodule was positive for adenocarcinoma, and RUL was non-diagnostic. Patient was diagnosed with stage IA adenocarcinoma (T1bN0M0). The thoracic surgeon ordered a blood-based biopsy that showed no targetable tumor drivers; however, a KRAS-activating mutation (a genetic marker for poor prognosis) was detected and proteomic analysis revealed a "poor" prognosis.
The patient was not a surgical candidate, and instead was treated with curative stereotactic body radiation therapy at the LLL site. Observation and monitoring were prescribed for the RUL. In February 2017, patient re-presented with disease progression. A new left hilar mass with mediastinal invasion was identified and determined positive for recurrence. Patient was hospitalized and placed on hospice shortly after recurrence was identified, less than a year after initial diagnosis.
Typically, patients presenting at this early stage have a 5-year survival rate of approximately 70 percent (Am Soc Clin Oncol Educ Book 2014:e375-380). While this patient's cancer was pathologically staged as early stage and potentially curable, the biology of the disease was aggressive and multiple prognostic markers (proteomic and genomic) showed a poor prognosis. The recognition of the effect of these markers in this case could help the multidisciplinary team build a more aggressive treatment plan for the next patient who presents with a similar prognosis in an effort to prolong survival.
Case Study 3
A "good" prognosis and response to chemotherapy
A 59-year-old female presented in February 2017 with pain in her shoulder; scans revealed a 6 x 4.5 cm mass in the LLL that was diagnosed as a stage II adenocarcinoma. A liquid biopsy was ordered that detected a KRAS G12V mutation, and proteomic analysis showed a "good" prognosis. The patient was a candidate for surgery; however, the presence of some chest wall invasion led the treating oncologist to prescribe neoadjuvant chemotherapy. Based on the prognosis and mutation status, cisplatin-based doublet chemotherapy was chosen, and the patient had a radiographic partial response. Twelve weeks after diagnosis, the patient underwent a lobectomy and lymph node dissection. A liquid biopsy was ordered after surgery and no KRAS mutation was detected. Based on the surgical outcome (node-negative and R0 resection) and absence of KRAS mutation after surgery, no adjuvant therapy was prescribed. The patient is under observation and currently shows no evidence of disease.
Inclusion of liquid biopsies during the initial pulmonology visit can provide important predictive and prognostic information during initial multidisciplinary team discussions. The rapid turnaround time from liquid biopsies allows for a much faster treatment start. The use of objective proteomic prognostic markers in conjunction with other clinical factors can help the medical oncologist determine how best to treat an individual patient and implement the treatment plan quickly without waiting for tissue biopsy results. Consistently receiving and using this type of objective information in a multidisciplinary setting can help all team members devise better and more personalized methods for patient care.
KRISH BHADRA, MD, is Interventional Pulmonologist at Lung Care Associates. SUSAN GARWOOD, MD, is Thoracic Oncology Lead at Sarah Cannon Institute. NITIKA SHARMA, MD, is Faculty at Eastern Carolina University, Greenville, N.C. PAUL WALKER, MD, FACP, is Division Chief at Eastern Carolina University.