Conference Report - Early Cancer Diagnosis: Beating the Odds

Introduction

What can be done to improve the early diagnosis of colorectal cancer (CRC), breast cancer (BC), and cervical cancer (CC)?

Experts convened at the 95th Annual Meeting of the American Association for Cancer Research, recently held in Orlando, Florida, and presented data supporting the efficacy of the preventive measures implemented so far. Prevention can indeed change the historical course of a disease in a population at risk, as shown by the results obtained for breast and cervical cancers. The social and demographic changes that will be occurring in the years ahead and the need to improve on these results are, however, calling for new strategies and renewed efforts.

New insight in the pathogenesis of cancer and novel technologies are offering the tools to improve on the prediction value of screening procedures and their applicability to large sections of a population, in view of the practical, social, and financial issues associated with their implementation. The focus here is a perspective on the old and new ways to screen for CRC, BC, and CC.

CRC Screening

CRC is the second leading cause of cancer mortality in the United States. It affects men and women equally: Men have a 1:17 lifetime risk of being diagnosed with CRC, whereas women have a lifetime risk of 1:18. The risk of death for both is about 1:42-1:43. Thus, the perception that some might have that this is a men's disease is quite unfounded.

As illustrated by Dr. Bernard Levin[1] of the MD Anderson Cancer Center, Houston, Texas, the benefits that derive from early diagnosis are clearly documented -- 90% of CRCs are curable if diagnosed at a localized stage, whereas only 9% of disseminated CRCs can be cured. A fecal occult blood test (FOBT) and a lower endoscopy at least every 10 years can lead to the diagnosis of 21.5% and 45% of cancers, respectively. When combined, their diagnostic potential is higher at 54%.

Overall, however, only approximately 40% to 50% of the individuals at risk have been screened for CRC by FOBT between 1997 and 2002[2-4] as compared with 80% to 90% of women at risk of BC or CC undergoing mammography or Papanicolaou (Pap) testing. The gap is still far too large for a large impact of early diagnosis on the natural history and increasing prevalence rate of CRC in developed nations. More awareness needs to be raised among the general population.[5]

Multidisciplinary task forces are, thus, advising more frequent testing to optimize the early diagnosis of CRC, as several methods are effective for screening of men and women over 50 years of age. The American Cancer Society advises an FOBT test every year and a flexible colonoscopy every 5 years. The US Preventive Services Task Force (USPTF) also supports a yearly FOBT for individuals at risk, although it leaves unspecified the optimal interval for CRC screening by flexible colonoscopy. Benefits have been shown to significantly outweigh the risks and costs.[1]

FOBT

FOBT is a noninvasive screening procedure, but it has limited effectiveness owing to a lower sensitivity. It does not require any preparation of the subjects and it can be readily performed by a general practitioner. Some patients may, however, find it distasteful. The test is based on fecal immunochemical testing (FIT) with the detection of globin in the feces. Specificity can be as high as 97%. A positive test is, in fact, highly associated with the presence of a tumor: In 89% of cases it signals a CRC and 77% of cases an advanced adenoma with the Insure test. Data reported in 1993 by Young and colleagues[6] showed a substantial impact of FOBT testing on the incidence and early diagnosis of CRC.

Optical Colonoscopy

Colonoscopy has the highest sensitivity, and it allows diagnosis and endoscopic treatment within the same session. It is usually performed under sedation and, thus, patients seem willing to accept it. It has been proven cost-effective and it yields protection with the longest interval. No controlled trial results are available, but a mild risk of perforation (although highly variable according to institution/physician) is associated with this procedure, and estimated at about 1:300-1:4,000 of individuals screened.[1]

DNA Testing

New procedures are being devised to optimize CRC screening and render it easy to implement in large populations of individuals at risk at reasonable costs. DNA testing for a set of genetic markers associated with CRC is being implemented.[3] Among the tumor markers included, there are the microsatellite-instability markers, Bat-26, the transforming growth factor receptor 2, Bax, the insulin growth factor receptor, and others.

From a practical standpoint, DNA testing is a noninvasive procedure that can be easily performed at home. It does not require any change in diet or premedication. Only 1 specimen is needed, thus avoiding the need to handle feces. Specificity for CRC detection with 1 of these kits (PreGen-Plus) is high at 95%, but sensitivity is significantly lower at 52% -- approximately half of the tumors still go undetected.[1] Studies are in progress to define the optimal interval between tests.

Virtual Colonoscopy

Virtual colonoscopy (VC), the latest technique proposed for screening of patients at risk of CRC, relies on helical computer tomography of thin sections, with the patient in the supine and prone positions.[3] Recently published results from 3 major centers[7] indicate a higher sensitivity for the detection of polyps in asymptomatic individuals when compared with conventional colonoscopy: 94% detection with virtual vs 88% to 92% with optical colonoscopy. Sizewise, 30% of the polyps referred were 6 mm, 14% were 8 mm, and 8% were 10 mm, supporting the high-resolution power of this diagnostic technique. Of note, extracolonic findings of high clinical importance were reported in 4.5% of patients._[7]

Such imaging technology yields huge data sets, and there is a steep learning curve for radiologists, but automation is being considered. Of note, visual colonoscopy misses flat adenomas, and the technology is being modified to overcome this anatomic limitation. Also, the possibility to perform a prepless VC is being addressed to avoid the currently needed oral preparation of the patient, with mild catharsis and mild diet.

Thus, in the future, both quantitative and visual reports will be provided, and more patients will be screened with an even higher number of polyps being detected. VC, in fact, can examine an entire colon independent of strictures, and it is believed to be twice as accurate as flexible sigmoidoscopy.[1]

Dr. Levine, however, acknowledged that, "not all the population over 50 years of age can be screened by VC in most countries." Thus, in general, a "funnel-like" approach may be the most effective strategy for large-scale CRC screening, with FIT as the initial step, followed by DNA testing, and then optical colonoscopy and/or computerized tomography.[1] New findings in the proteomic profiles of adenomas may render FITs far more sensitive in the future, and thus more effective in the screening of premalignant and early malignant lesions.

Biomarkers for Early BC Diagnosis

The identification of tumors while they are still localized and curable leads to significant reductions in overall mortality and morbidity. The advent of new technologies, such as gene expression microarrays, genomic analysis, and proteosome analysis, are providing the tools to make preventive approaches even more effective than in the past, both in terms of early diagnosis and in the prediction of clinical outcomes.

A number of factors are known to contribute to the mortality and morbidity rates in cancer patients: Health behaviors (eg, smoking in lung cancer), 40%;

• Genetic factors, 30%;

• Socioeconomic circumstances (eg, lack of access to medical care for mammographies), 15%;

• Quality of medical care, 10%[8]; and

• Environmental exposure to oncogenic compounds, 5%.

But among BCs, how many are of predominant genetic origin? Approximately, 5% to 10% of BCs are hereditary and 15% to 20% occur in family clusters. In the latter case, environmental and cultural factors may coexist with genetic factors, inducing higher rates of BC in women of the same family. Most of BCs appear to occur in a sporadic fashion. BC may occur in men, although with a very low incidence rate.

As illustrated by Dr. Olufunmilayo Olopade[9] of University of Chicago Medical Center, Chicago, Illinois, DNA repair pathways play a significant role in BC susceptibility, particularly in patients with defects in the BRCA-1 and BRCA-2 genes.[10-12] Two other molecules, ATM and Check2, are essential for a correct functioning of the intracellular pathways activated in response to DNA damage.[13]

Approximately 30% to 40% of genetic BC involves defects in BRCA-1, 20% to 30% in BRCA-2, and only a minority of mutations in PTEN or in ATM/Check2.[14-16] An association has been found between the dysregulation of PTEN and BC in patients with Cowden's syndrome, who also may develop thyroid carcinoma and tubular adenomas in the bowel. The specific mechanisms underlying the remaining 30% to 40% of genetic BC have not yet been identified.

Often when mammographies are performed in 35-year-old women, high tissue density may generate diagnostic uncertainties as to the presence of potentially malignant lesions. Nuclear magnetic resonance imaging may be of help in these cases to avoid subjecting high-risk women to repeated biopsies in the course of their lives.

Computerized cancer-risk assessment may also represent a new way to define individual risk of BC in a more quantitative way. Studies have been performed to compare mtBRCA-1/-2 carriers vs low-risk, age-matched groups. Texture features (including the percentage of density and texture of parenchyma) were computer-extracted from original mammograms and then assigned a numerical value related to the risk of BC. A coarse parenchymal pattern with low contrast was identified more frequently in women at high risk of BC.[17]

Hereditary BC associated with mutations in BRCA-1 and BRCA-2 is more frequently of higher histologic grade (2 and 3) and negative for expression of the estrogen and progesterone receptors.[18] Only a minority, in fact, are hormone-receptor positive (about 20%).[16] The oncogene HER-2/neu is not frequently amplified in mtBRCA-1-associated cases, whereas the oncogene c-myc has been found overexpressed in basal-like tumors independent of BRCA status. Amplification of chromosome 8 and monosomy of chromosome 17 can be detected in both hereditary and sporadic BC, whereas aneuploidy and other chromosomal aberrations and amplifications are observed more often in women carrying mutant BRCA-1.[9]

Cell-cycle components as p27 and cyclin D1 are downregulated in mtBRCA-1 tumors, but only p27 in mtBRCA-2 tumors. Conversely, cyclin D1 is increased in lesions positive for mtBRCA-2 and cyclin E is increased in mtBRCA1 tumors. mtBRCA-1 tumors are also positive for the antiapoptotic factor Bcl-2 and the cell proliferation marker Ki67.[9]

BRCA-1 mutations can be found in stem cells or transitional elements, which are characterized by high proliferation rates. Thus, according to 1 model, BC is thought to proceed from clonal expansion of susceptible cells that usually have basal-cell identity. A second model envisions the loss of BRCA-1 function in more mature epithelial cells with the loss of a critical checkpoint necessary for the maintenance of cell homeostasis following DNA damage.

A number of chemoprevention approaches are being investigated, including hormone antagonism, aromatase inhibition, and cyclooxygenase inhibition. More insight in the mechanisms that generate and sustain the growth of sporadic and hereditary BC will allow the definition of effective and safe prevention strategies for women who live with a life-long, high risk of developing BC.

Human Papilloma Virus and Early Detection of CC

Cytologic screening of women at risk has led to a reduction in the incidence of cervical carcinoma of more than 80%. Hence, we can say that cytology works.[19] Less than 5000 deaths occur annually in the United States and, of note, half of these deaths occur in underscreened women. Incidence and death rates by CC are, however, far higher in many other countries where women have little or no access to preventive healthcare.

Nonetheless, approximately 50-60 million Pap tests are done each year in the United States, 3.5 million of which require further evaluation -- quite a burden for the healthcare system. Is there an adjunct to screening that might help in triaging women with minor abnormalities, or even a replacement of conventional cytology that would allow identification of those women that require further treatment or follow-up?

As illustrated by Dr. Thomas C. Wright[19] of Columbia University, New York, New York. a number of potential biomarkers are being investigated to this end in premalignant and malignant CC lesions:

• HPV with DNA-, RNA-, or protein-based tests;

• Cell-cycle proteins (eg, p16 INKa);

• Other markers identified by transcription profiling (eg, fibronectin, metalloprotease-9); and

• Markers derived by spectroscopic analysis.

More than 100 different HPVs are now being identified and classified according to their species specificity. They are epithelium-tropic viruses that cause proliferation of infected cells. Almost all CCs have been found to contain HPV sequences, particularly those linked with a high risk of CC (eg, HPV-16 and HPV-18). The E6/E7 oncoproteins are critical in inducing the transformation of HPV-infected cells in a variety of systems.[20]

Pap tests can detect mild infections by documenting an acute cytopathic effect; 10% of these cases go on to develop persistent infections. More than 2 million women in the United States are diagnosed with abnormalities in the cervix, following repeated Pap testing.[21,22] Six recent studies have, however, found that HPV testing led to the identification of more women with CIN2 and CIN3 lesions than repeated cytopathology. Molecular tests can now be easily performed after performing cytologic analysis by spinning down the liquid Pap solution and performing molecular analysis by reverse-transcriptase polymerase chain reaction (RT-PCR) or hybrid capture assays on cells in the pellet. "HPV testing is becoming the standard of care in the US," said Dr. Wright.[19,23]

When should women be screened for CC? HPV infection rates are high in very young women and they decrease after 35 years of age. CC rates, on the other hand, are very low at a young age and then they start to increase at age 30-35. Thus, screening for CC should start after 30. Sensitivity in detecting CIN3 lesions by Pap testing is quite variable, ranging from 52% to 90%. HPV testing is more sensitive with a yield of 71% to 100%. If both techniques are applied, sensitivity reaches 90% to 100%.

In the prospective study performed by the National Cancer Institute group led by Sherman, women underwent cervical/vaginal lavage, and frozen samples tested for the presence of cellular abnormalities. Approximately 4% of the HPV-positive women had developed CIN3 lesions after 3 years, and 7% of them at 10 years. Conversely, almost no lesions were observed in HPV-negative women.

On the basis of the results obtained so far, the interim US guidance says that women over 30 if negative at first HPV screening should be rescreened 3 years later.[19] Europeans, on the other hand, advise undertaking a second screening 5 years later, if the first result is negative.

But how should the 1,875,000 women found to be HPV-positive in the United States be managed? Both Pap and HPV testing should be repeated in 6-12 months. Those negative for both should be retested 12 months later. Conversely, women positive for high-risk HPV and/or with a positive Pap cytology should be referred for colposcopy.[19,23]

In answer to the question from the audience -- "Why should we spend time and money screening instead of vaccinating everybody?" -- Dr. Wright said, "the vaccine, even if 100% effective, would be only prophylactic and not therapeutic. All women infected in the past by HPV should still be screened for CC. Thirty years will be needed to see the impact of a successful vaccine worldwide." And that will be a day.