John B. Kisiel, MD, and Philip Parks, MD, MPH, provide an update on the future of cancer screenings and explain how improved treatment regimens offer the potential to reduce the cancer-related healthcare cost burden.
Cancer remains the second leading cause of death in the United States.1 As 2018 comes to a close, an estimated 1,735,350 new cancer cases will have been diagnosed and 609,640 individuals will have died from the disease.2 Although cancer therapies continue to improve outcomes, more effective screening technolgies— enabling more people to get screened and more cancers detected at an earlier, more curable stage– are needed. Better and more convenient screening tools for a variety of cancer types will help us achieve the targets set forth by leading healthcare organizations. Screening for common cancers is generally cost effective,3-5 and an increase in screening rates accompanied by earlier detection of cancer and improved treatment regimens offer the potential to reduce the cancer-related healthcare cost burden through improving outcomes and survival.6-7
Colorectal Cancer Screening
As we look optimistically to a future state with more than 80% of the population screened for colorectal cancer (CRC),8 we are also faced with 2 primary challenges. First, although CRC screening rates are modestly improving,9 they have not yet reached stated goals, despite national campaigns led by influential organizations including the American Cancer Society (ACS), National Colorectal Cancer Roundtable (NCCRT), and large integrated delivery systems and advocacy groups. The barriers to screening include factors related to patients, healthcare providers, health systems, and communities. These barriers are difficult to overcome even though CRC screening reduces the incidence of CRC by one-half and mortality by one-third.2,10 Second, evolving epidemiologic evidence demonstrates that there is a disturbing “birth cohort” effect that highlights a 51% increase in the incidence of CRC from 1994 to 2014 and an 11% increase in mortality from 2005 to 2015 among individuals 55 years and younger.11,12 Based on these incidence and mortality data, the ACS has recommended that screening begin earlier, at age 45, for average-risk individuals.
With a goal of 80% of the average-risk population screened according to guidelines, clinicians, population health specialists, payers, employers, and large integrated health networks must fundamentally change and improve CRC screening programs. Improvements in CRC and other cancer screening programs aligns with the Triple Aim of healthcare: (1) better, higher quality of care; (2) healthier populations and communities; and (3) more affordable care.13 One positive pivotal change that occurred in health policy was for the US Preventive Services Task Force (USPSTF) to provide an “A” rating for CRC screening for individuals aged 50 to 75 at average risk for CRC.14 With their recommendation for any 1 of 7 screening strategies in the 2016 USPSTF update, patients and providers are encouraged to “choose the best test that gets [it] done.” The NCCRT, established in 1997 by the Centers for Disease Control and Prevention and ACS, developed webinars, handbooks, and other resources for hospitals and health systems to support implementation of best practices in CRC.15
A high-quality screening test must have 3 characteristics: (1) high sensitivity, (2) compliance and adherence, and (3) access via insurance and shared decision making. Patient values and preferences play a significant role in compliance and adherence.16,17 The majority of patients in the United States continue to be screened with colonoscopy,9 which requires bowel preparation, time away from work, sedation/anesthesia, and risk of complications from preparation, sedation, or the procedure.18-20 Many patients are apprehensive, if not fearful, about screening21 and may prefer a noninvasive stool-based screening test.16 According to the 2018 ACS guideline update, “Although prevention is highly valued by patients, test preparation, invasiveness, potential costs, and other considerations will lead some patients to prefer a noncolonoscopy test for screening.”16
Currently there are several noninvasive tests, including the fecal immunochemical test (FIT), the guaiac fecal occult blood test, and the multitarget stool DNA, which are all included in the USPSTF recommendations.14 Stool-based tests vary in sensitivity, with the multitarget stool DNA having the highest sensitivity for CRC (92% vs the FIT test’s 74% sensitivity in a head-to-head trial),22 and the highest sensitivity for adenomatous and sessile serrated precursors.22 In addition to test performance and considerations of patient preference, another evidence-based component of successful screening programs is patient navigation.23-25 Patient navigation programs vary in scope and effectiveness by hospital and health systems; however, 1 screening strategy, the multitarget stool DNA test, includes an embedded nationwide patient navigation program.26
There is also now 1 blood-based screening test for CRC that is available by FDA label to patients who are unwilling or unable to screen with other recommended CRC screening choices in the 2016 USPSTF guidelines. The Epi proColon test detects methylated Septin-9 and has a sensitivity of 68% and specificity of 80%.27-29 Although liquid biopsy for CRC screening is desired, current scientific and engineering limitations delay the availability of a blood-based screening test for precancer and cancer with sensitivity and specificity as high as stool-based testing. There are numerous research and development efforts underway to improve the molecular technology necessary to commercialize an effective blood-based screening test for CRC. The future of CRC is one where patients and providers make choices based on patient values and preferences, clinical performance, and awareness of risks, benefits, and alternatives, with the inclusion of patient navigation systems to optimize compliance and adherence.
Lung Cancer Screening
Because molecular testing for early detection of CRC has made such inroads into the clinical space, we and others anticipate that similar chemistry could be applied to screening for the leading cancer killers. Of these, lung cancer alone accounts for 25% of all cancer deaths, with a loss of 154,000 lives annually.30 To date, the only effective screening option for lung cancer is a low-dose computed tomography (LDCT) scan. The National Lung Screening Trial (NLST) involved over 53,000 current or former smokers, who were randomized between screening with LDCT versus chest x-ray. After 3 screens, there was a 20.0% reduction in lung cancer deaths.31 Based on these results, the USPSTF recommends LDCT screening for those aged 55 to 80 with a 30 pack-year history of smoking and who either still smoke or have quit within 15 years.32 Endorsement of a screening strategy represents a major leap forward; however, this approach is currently applied to too few at risk. A study based on the findings of the NLST found that if computed tomography (CT) screening was implemented among screening-eligible US populations, only 12,250 deaths, fewer than 10% of the current annual lung cancer mortality, would be averted each year.33 Too few of those who are at high risk for lung cancer do not fit the recommended criteria; especially those who quit smoking more than 15 years ago. Additionally, lung cancer screening by CT may result in unnecessary intervention due to a false-positive rate that exceeds 96%.31
A blood-based screening test may offer superior clinical performance and improve access for patients. One promising approach entailed examining DNA methylation patterns using next-generation DNA sequencing in primary lung tumors and high-risk control tissues to identify highly sensitive and specific methylated DNA markers (MDMs) of lung cancer. These MDMs were validated in DNA extracted from independent tissue samples and clinically validated in archival (EDTA)-buffered plasma specimens from 23 cases and 80 controls. Early results showed that a 4-MDM panel achieved an overall sensitivity of 96% and specificity of 94%.34 The MDM panel is currently being optimized for improved sensitivity of early stage lung cancers, and prospective enrollment of a phase 2 validation study is currently in progress. The future of lung cancer screening is promising as molecular techniques pave the way for a more accurate and convenient blood-based test.
Liver Cancer Screening
Lung cancer screening by blood-based assays is biologically rational due to circulatory anatomy that gives tumor-specific DNA access to the plasma compartment; 100% of cardiac blood output passes through the lung. Another organ that receives high cardiac output is the liver, which outflows directly into systemic venous circulation. Although substantially less common than lung cancer in the general US population, primary cancers of the liver (hepatocellular carcinomas [HCCs]) are the second leading cause of cancer deaths worldwide and are projected to be the fourth leading cause of cancer death in the United States by 2030.35 HCC primarily arises in patients with chronic liver disease. The most common of these are hepatitis B and C infections, alcoholic hepatitis, and non-alcoholic fatty liver disease (NAFLD). Because of the obesity epidemic, NAFLD is the most rapidly increasing risk factor for HCC.
Surveillance for HCC is supported by the results of a randomized controlled trial in patients with hepatitis B in China, which demonstrated a near 40% reduction in mortality among those surveilled by ultrasound and serum assay of alpha-fetoprotein (AFP).36 While this trial has not been replicated in the West or by using patients with other liver diseases, there is ample observational data that patients under surveillance are more likely to be diagnosed with HCC at earlier stages, which may improve survival. The main drawback to this approach is low sensitivity for curable-stage disease. A recent meta-analysis estimates that ultrasound and AFP in combination are only 63% sensitive for early-stage HCC.37 Moreover, adherence to surveillance testing is quite poor.38
In a similar approach to the discovery of lung cancer markers described earlier, DNA from primary HCC tumors and control liver tissues was sequenced to identify MDMs associated with HCC. The candidate MDMs were validated in independent samples before pilot testing in archival plasmas of 21 patients with HCC and 30 patients in the cirrhosis control group. A 2-marker MDM panel was found to be 89% sensitive and 87% specific in the pilot phase of clinical testing.39 A larger phase 2 study assayed MDMs from archival plasma samples of 95 HCC cases, 51 cirrhotic controls, and 98 healthy controls, and a 6-marker MDM panel was found to be 95% sensitive for HCC at 92% specificity.40 Most importantly, 93% of HCC tumors that were of curable stage were detected at the same specificity threshold. Larger phase 2 and phase 3 studies are in progress to set strict cut-offs for MDM markers in a clinical assay and determine if MDMs can detect HCC prior to other surveillance modalities. As with lung cancer screening test development, advances in liver cancer screening using a blood-based test show promise.
Multicancer Screening: The Path Forward
For lung and liver cancer, surveillance is targeted to high-risk patient subsets where the prevalence of cancers is enriched by a predisposing chronic illness. Yet most cancer deaths occur in persons without a known predisposition. Unfortunately, population screening has not been justified for most other cancers due primarily to individual prevalence rates that are insufficient to allow cost-effective interventions. At a population-wide level, benefits of single-organ screening have been demonstrated most robustly for breast, cervix, colorectum, and, to some extent, prostate cancer. The relatively high prevalence of these cancers directly affects the positive predictive value of screen testing and the number of patients needed to screen to identify a specific cancer. For instance, roughly 1 CRC will be found among 170 persons screened. However, 500 to 1000 persons would need to be screened to identify pancreatic or esophageal adenocarcinomas, respectively, due to lower prevalence of these very fatal diseases. As a result of screening so many persons, even a very specific test is anticipated to generate an unacceptable number of false-positive results, resulting in expensive downstream testing and unnecessary patient anxiety.
A screening test capable of detecting multiple cancer types is an attractive option to fill this gap. If lower and higher prevalence cancers could be screened simultaneously, the combined prevalence in the screened population would dramatically lower the number needed to screen.41
Next-generation DNA sequencing and other emerging technology platforms are being leveraged to identify markers that appear to have high sensitivity and specificity for cancers and appear to identify patterns predictive of the anatomic origin of the primary tumor. Our group has demonstrated proof-of-concept data that multiple cancer types can be detected from the same biological media, including blood and stool, using MDM assays directed towards markers of both pancreatic cancer and CRC.42 Other investigative teams have combined DNA mutation and protein markers to detect multiple cancer types and have developed data models that associate biomarker patterns with each primary cancer type.43 These observations herald an exciting and potentially transformative new direction in the fundamental approach to cancer screening. Noninvasive multicancer screening is expected to be a near-term reality and will fuel rapidly increasing competition in research and commercialization efforts to bring this concept to clinical practice.AUTHOR INFORMATION
John B. Kisiel, MD, is a consultant in the Division of Gastroenterology and Hepatology, Department of Internal Medicine, and an associate professor of medicine at the Mayo Clinic. His research focus is the prevention and early diagnosis of cancers and precancers in high-risk patients, including those with inflammatory bowel disease.
Philip Parks, MD, MPH, is the senior director of medical affairs, Exact Sciences. He is a board-certified practicing physician with more than 16 years of clinical, corporate medicine, and managed care experience. Dr Parks served in the US Navy as a diving and undersea medical officer and held technology assessment leadership roles.
DISCLOSURE. Mayo Clinic and Exact Sciences jointly own intellectual property on which Dr Kisiel is listed as an inventor and may receive royalties.References: