BBI Contributing Editor

ORLANDO, Florida The 41st annual meeting of the American Society of Clinical Oncology (ASCO; Alexandria, Virginia), held here in mid-May, provided a forum for presentation of the latest developments in cancer diagnosis and treatment worldwide, and included announcements of some important advances in drug therapy for major cancers. A major trend is now well established for the use of targeted drug strategies to provide more effective treatment for specific subsets of patients while reducing the incidence of side effects that plague conventional chemotherapy regimens. Targeted therapy relies on identification of specific molecular entities such as growth factors or receptors that play a key role in tumor proliferation or cancer cell metabolism, and developing drugs that attack those entities to selectively kill tumor cells or halt tumor progression. Most new drugs now being developed for cancer therapy are targeted agents, and advances in treatment are increasingly becoming dependent on the use of targeting strategies.

Those developments have implications not only for pharmaceutical and biotechnology companies involved in cancer drug development, but for diagnostics companies as well, since there is growing demand for tests to identify patients who will respond to targeted drugs. Genetic testing also is playing an increasingly important role in cancer management, as new discoveries in genomics are providing tools that not only allow improved selection of treatment in patients diagnosed with cancer, but perhaps more importantly are allowing individuals at high risk for cancer to be identified so that disease can be detected at an early stage when it is most curable.

Furthermore, a number of presentations at this year's ASCO conference focused on cancer prevention, an area that can benefit greatly from the use of technologies that are able to determine which individuals will benefit most from preventative strategies. A wide range of technologies for early detection of cancer was described at the conference, including advances in cellular analysis technologies as well as diagnostic imaging. New developments in device-based cancer therapy also were described at the conference, including treatments that combine radiation therapy with chemotherapy to enhance therapeutic effect, as well as modalities such as hyperthermia for the treatment of advanced cancers.

In spite of the advances that have been made in cancer diagnosis and treatment over the past decade, the impact of cancer on public health continues to increase, as cancer is now the leading cause of death in the U.S. for individuals under the age of 85. The aging of the population in the developed world virtually guarantees that cancer will become an even more important disease in the future, since cancer incidence increases steadily with age, peaking at around age 85. As a result, the market opportunity for new technologies for cancer diagnosis and treatment is expected to continue to expand.

Explosion in demand for targeted therapy

Targeted cancer drugs now account for a substantial portion of the total cancer drug market worldwide. Genentech (South San Francisco, California), the leading supplier of targeted cancer agents, reported worldwide sales of $2.8 billion for its bio-oncology drug portfolio in 2004, up 44% vs. 2003. The products include Rituxan, a drug that targets the CD20 cell surface marker for treatment of non-Hodgkins lymphoma; Avastin, which targets vascular endothelial growth factor (VEGF) to cut off a tumor's blood supply, and is used in the treatment of metastatic colorectal cancer; Herceptin, which targets the her-2/neu receptor that is amplified in certain aggressive forms of breast cancer; and Tarceva, a drug for the treatment of non-small cell lung cancer which targets the epidermal growth factor receptor (EGFR) that is believed to play a key role in the tyrosine kinase signaling pathway in tumor cells responsible for cell growth.

In the case of Herceptin, the first targeted agent to be developed, pre-screening of patients with a molecular or immunohistochemistry test to determine HER2 positivity is mandatory to qualify a patient for treatment. The Herceptin example, combining a gene expression test with a companion targeted drug to improve treatment efficacy, is also an example of one of the first widespread uses of pharmacogenetic testing. In principle, the use of other targeted drugs is also directed by the results of such tests, in order to verify that the drug's target is present in the patient. However, that has not always proven to be the case, since patient response is not always accurately predicted by testing, as is the case, for example, with drugs that target EGFR. In that example, only 80% of patients who respond to the drug are EGFR-positive by diagnostic testing.

Nevertheless, the future course of cancer drug development, a field in which 1,500 new agents are now in various stages of development, has become firmly established on the principle of identifying molecular targets that can be attacked with specific agents. While experts presenting at the ASCO conference predict that cytotoxic chemotherapy drugs will still be the primary modality used five years from now, they expect a continued increase in the use of targeted therapies in the future. From a market perspective, the trend is likely to drive strong growth in dollar volume sales, since new targeted agents command significantly higher prices than older chemo-therapy drugs. Avastin, for example, costs $4,000 for a two-week treatment course.

The task of developing targeted drugs to treat all types of cancer is a daunting one, however, since cancer is not one but about 700 different diseases, each of which may have different underlying molecular mechanisms. Developers have thus focused on drugs that target major cancer types, which include breast, lung, and colorectal cancer, as well as non-Hodgkins lymphoma, as shown in Table 3. Those cancers account for about 43% of cancer patients in the U.S., although the percentage of patients who are candidates for treatment with existing targeted drugs is smaller since not all patients express the target marker. Furthermore, none of the targeted drugs have proven to be a complete cure, since resistance mechanisms eventually develop in most patients that allow the tumor cells to circumvent the targeted pathway.

In some cases, targeted drugs are combined with conventional chemotherapy agents. For example, studies presented at the ASCO conference found that Avastin combined with paclitaxel and carboplatin improves overall survival in patients with advanced non-small cell lung cancer by 30%. Avastin is a particularly promising targeted agent because its spectrum of action is potentially quite broad, extending in principle to any solid tumor that is dependent on blood vessel growth (angiogenesis).

Another cancer in which the drug has shown efficacy is metastatic breast cancer, where an interim analysis of a study presented at the ASCO conference showed a 49% improvement in overall survival and a response rate of 28% vs. 14% in patients receiving chemotherapy alone, and studies are in progress for pancreatic, ovarian, renal cell and other cancers.

Promising results also were announced at ASCO for a new targeted agent for the treatment of advanced kidney cancer, AG-013736, being developed by Pfizer (New York). The drug targets VEGF and platelet-derived growth factor (PDGF) receptors, which are strongly expressed on blood vessels in kidney tumors. The Pfizer drug is also being evaluated in breast cancer, non-small-cell lung cancer, melanoma, and thyroid cancer. In a study of 52 patients with cytokine-refractory, metastatic renal cell cancer described by Brian Rini, MD, of the University of California, San Francisco, an 86% response rate was achieved, including 46% with partial response and 40% with stable disease. Side effects were mild and included fatigue, hypertension, nausea and diarrhea.

Targeting of oncogenes is another strategy for development of new cancer therapies that is conceptually attractive and showing promise in a number of cases. As discussed by Harold Varmus, MD, of Memorial Sloan Kettering Cancer Center (New York), oncogenes that play key roles in cancer include abl, erb-b, c-myc, c-kit, p53, raf-1 (an angiogenic factor), and src. Two types of oncogenes may prove to be important targets, including those that initiate cancer and those that maintain growth of a tumor once it has developed, with raf-1 being an example of the latter type.

As discussed by Dean Felsher, MD, PhD, of Stanford University School of Medicine (Palo Alto, California) at the ASCO conference, c-myc is one of the most attractive oncogene targets because it is involved in the regulation of tens of thousands of genes. It is closely involved with lymphoid malignancies in particular. Felsher has demonstrated the ability to inactivate c-myc in animal studies with doxycycline, a broad-spectrum anti-infective drug. Inactivation results in reversion of tumor cells to normal cells, or at least to a non-proliferative state, and reversal of inactivation does not necessarily result in restoration of tumor growth.

The effect of oncogene inactivation is dependent on the tissue environment. For example, c-myc inactivation in liver tumors results in conversion of the tumor cells to stem cells, which can then revert to tumor cells when the oncogene is reactivated. Felsher is now using DNA microarrays to help in assessing the complex interactions of oncogenes in order to predict when inactivation of certain oncogenes will be effective. Another target is telomerase, which is involved in control of the length of telomeres in the cell. As discussed by Steven Artandi, MD, of Stanford University, telomere shortening is a hallmark of aging, and is even more pronounced in malignancy.

The growing importance of targeted therapeutics is creating strong interest in development of pharmacogenetic tests and related methods for therapy guidance. William Hahn, MD, of Dana Farber Cancer Institute (Boston), is now developing oncogene microarrays that will be used to screen for drug effects. Because of the complexity of oncogene expression patterns, microarrays are likely to play a major role in helping to determine the results of targeted drug treatment, at least in the discovery phase.

Physicians are beginning to move beyond anatomic staging of a tumor to the use of gene tests to provide a more specific characterization of the tumor that is more closely linked to drug response. In breast cancer, for example, at least three genes, the BRCA 1 and 2 genes, as well as HER2, are important for guiding therapy. Gene expression profiling studies in breast cancer described by Debu Tripathy, MD, of the University of Texas Southwestern Medical Center (Dallas), are using the Oncotype DX21 gene expression test from Genomic Health (Redwood City, California). Evaluations of breast cancer patients using the Oncotype DX test have demonstrated distinct clusters of gene expression that correlate with response to therapy. In a validation study conducted with the assay, 47% of breast cancer patients were reclassified using the Oncotype DX when compared to National Comprehensive Cancer Network (NCCN) guidelines, and use of the test allowed more accurate prediction of breast cancer recurrence. In particular, Oncotype DX reclassified 49% of NCCN high-risk patients, a group that comprised 92% of all patients in the study, to low risk, thus potentially sparing the use of unnecessary toxic chemotherapy and also lowering costs. Some 28% of NCCN low-risk patients were reclassified to intermediate or high risk, indicating that a different treatment strategy was required in order to avoid recurrence. The test is performed at Genomic Health's reference laboratory, with a turnaround time of about two weeks.

Tripathy also discussed the use of genetic markers used in addition to the HER2 oncogene marker in breast cancer. Other markers now proving useful in predicting response to Herceptin therapy include PTEN, phosphoHER2, and IGF-1R. PTEN is absent in 40% of HER2-positive cancers, but is required for activity of Herceptin. A study discussed by Tripathy found a 66% response rate for patients having high PTEN expression versus 11% for those with low expression. Tripathy believes more comprehensive gene profiling may prove useful in dealing with Herceptin resistance, by allowing the complex pathways involved in HER2-related breast cancer to be elucidated.

Gene expression profiling also has utility in predicting response to cytotoxic chemotherapy and hormonal therapy in breast cancer. As a result, management of breast cancer, as well as of many other cancers, is expected to move beyond the use of single markers to determine treatment to the use of gene arrays or multi-marker genetic test panels to allow simultaneous assessment of multiple markers.

Another important advance in technologies useful for guidance of breast cancer therapy described at the ASCO conference uses analysis of DNA methylation for prediction of the need for chemotherapy in node-negative breast cancer. Presently, most breast cancer patients receive chemotherapy following initial surgical treatment in order to eradicate tumor cells that may not have been detected and removed. However, it is known that a significant proportion of patients will have good outcomes, with greater than 90% event-free survival long-term, even if they do not receive adjuvant chemotherapy. The challenge is to identify those good-prognosis patients so that they can be spared exposure to the toxic side effects and costs of unnecessary treatment.

As discussed by Nadia Harbeck, MD, of Technical University of Munich (Munich, Germany), methylation of the PITX2 gene as measured by real-time PCR, a technology marketed by Roche Molecular Diagnostics (Pleasanton, California), is highly predictive of recurrence, with 98% of patients having low levels of PITX2 methylation metastasis-free at 10 years compared to 85% in the group having high methylation levels. DNA methylation has been proposed by a number of researchers as an important parameter in cancer development, due to its effect on gene function. Harbeck stated that a follow-up retrospective validation is already under way using the DNA methylation assay, and that if positive that study would be followed by a prospective clinical trial.

Another group, led by Christos Sotiriou, MD, of Bordet Institute (Brussels, Belgium), has used GeneChip arrays from Affymetrix (Santa Clara, California) for gene expression profiling of breast cancer. Previous studies have shown that there is poor agreement among different institutions in the histopathological grading of breast tumors, and in particular in Grade 2 cancers. Using a 22,283-probe array, Sotiriou created a Gene Grade Index that identified two strong patterns of gene expression associated with Grade 1 and Grade 3 tumors, and also showed an equivalent discrimination of patient prognosis as that obtained from histological grade. In addition, Grade 2 tumors by histology were found by gene expression to actually be a mixture of Grade 1 and Grade 3 tumors by gene expression, effectively allowing Grade 2 to be eliminated as a tumor classification parameter.

Sotiriou believes gene expression profiling can be used in place of histological grade to provide more accurate and reproducible typing of breast cancers, particularly since the study also found good reproducibility when the gene expression index was analyzed using different microarray platforms. While such an approach will need to be validated before it can be widely implemented in routine clinical pathology, the results indicate that gene expression analysis could eventually become the preferred technology for grading and typing of tumor specimens, although the technology platform that will prove most viable has yet to be determined.

Other suppliers of microarrays for gene expression analysis exhibiting at the ASCO conference include Illumina (San Diego), which recently completed its acquisition of Cyvera, including a lower-cost microarray technology that will have applications in in vitro diagnostics. Illumina has demonstrated a high degree of correlation between its microarray assays and assays performed with real-time PCR, as used in the DNA methylation assay described by Harbeck. The Molecular Profiling Institute, a unit of TGEN (Phoenix), performs gene-profiling tests using microarrays from Agilent Technologies (Palo Alto, California), including the Mammaprint assay for gene profiling in breast cancer specimens. The Mammaprint assay is performed as a reference lab test at a cost of $3,200.

The Molecular Profiling Institute also is offering the Target Now gene profiling assay, a microarray assay used to determine if one or more drug targets exist in their cancer tissue. The Target Now test is aimed at the 600,000 patients annually in the U.S. who fail to respond to conventional chemotherapy. It can, for example, identify patients with cancers other than breast cancer who are positive for the HER2 marker targeted by Herceptin. In one case cited by the company, a patient with pancreatic cancer tested positive for HER2 using the Target Now assay, and subsequently responded to Herceptin treatment.

Yet another approach for molecular profiling of cancer relies on proteomic analysis to identify expression patterns in tumors that allows them to be distinguished from normal tissue and to assess prognosis. One of the technologies for proteomic analysis uses mass spectrometry (MS) to create a protein profile that can be analyzed using pattern recognition techniques to identify specific patterns associated with outcome. As described by David Carbone, MD, PhD, of Vanderbilt University (Nashville, Tennessee), use of MS analysis for serum profiling in lung cancer has been shown to accurately predict prognosis, and to allow early detection in some instances. Carbone described a case in which three of four patients who were cancer-negative when screened using clinical criteria but tested positive via MS profiling developed cancer six months later.

Profiling is used now as a research tool to identify specific protein markers that are useful in cancer diagnosis and prognosis, but for clinical applications it may prove more effective to develop a more limited panel of tests for the markers shown to have diagnostic utility. In Carbone's study, two markers were identified from the protein profile, thymosin B4 and ubiquitin, which appear to be the key factors correlated with predictive power of the assay.

MS technology also has been used in a mapping mode, by using a tumor tissue section as the specimen, and scanning across the specimen in an imaging mode while analyzing the protein profile at each grid position in the image. That approach is effectively an ultra-high resolution alternative to immunohistochemical analysis and also avoids the need for use of staining procedures to identify protein markers. Inter-laboratory variability of MS analysis is, however, an issue for widespread implementation of the technology in clinical diagnostics.

Other examples of genetic and proteomic markers that are being evaluated for use in cancer diagnosis, prognosis, and therapy guidance include assessment of chromosome 11 loss in neuroblastoma, analysis of KIT protein mutations in gastrointestinal stromal tumors (GIST), and analysis of the 5Q31 deletion in patients with myelodysplastic syndrome. Edward Attiyeh, MD, of Children's Hospital of Philadelphia, described results of a study of 11q gene deletions in children with neuroblastoma as a method to discriminate patients having a poor prognosis. Neuroblastoma usually occurs in children under five years of age, and many patients undergo spontaneous remission. However, about half of all patients have metastasis upon initial diagnosis and have a poor outcome. Attiyeh found that 17% of patients with neuroblastoma have an unbalanced loss of 11q, and that those patients have a significantly worse outcome. The n-myc oncogene has been used for two decades as a prognostic marker in neuroblastoma, but the new marker provides additional prognostic value. About 800 to 1,000 cases of neuroblastoma occur annually in the U.S.

Michael Heinrich, MD, of Oregon Health Sciences University Cancer Institute (Portland, Oregon), has assessed mutations in exon 11 of the KIT gene in GIST patients and found a correlation between response to Gleevec, a targeted drug manufactured by Novartis Pharmaceuticals (Basel, Switzerland), and the presence of KIT exon 11 mutations. The exon 11 test could benefit about 10% of the 5,000 patients with GIST each year in the U.S. who are misdiagnosed as having another disease but have GIST and who would benefit from targeted therapy. In addition, those without an exon 11 mutation could be followed more closely to catch recurrence earlier.

The use of 5q31 mutations in patients with myelodysplastic syndrome (MDS), the most common cancer of the blood in adults affecting 40,000 to 50,000 in the U.S. annually, could help identify patients who require more intensive ongoing treatment to prevent recurrence. At present, therapy is usually scaled back after giving an initial high dose, but in 18% to 21% of patients with MDS the disease recurs. The 5q31 test could help identify those patients who would benefit from ongoing high-dose treatment with the drug lenalidomide (RevLimid), marketed by Celgene (Summit, New Jersey).

In addition to established suppliers of microarrays and genetic testing technologies used for gene profiling in cancer, two other companies announced new testing products for use in testing for therapeutic targets in cancer patients. Sysmex (Kobe, Japan) exhibited the OSNA system, a one-step nucleic acid amplification system under development for analysis of mRNA expression. The OSNA method avoids the need for an extraction step, and allows an amplification assay to be performed in as little as 30 minutes. The initial application being developed in cancer diagnostics is detection of lymph node metastasis in breast cancer sentinel node biopsy procedures.

Sysmex also is developing a proteomic analysis system using a protein chip format. The C2P analyzer measures multiple protein markers, and initially has been designed for analysis of cyclin-dependent kinase activity in cancer tissue specimens. The results of the test can be used to predict which breast cancer patients will respond to paclitaxel therapy.

Genzyme Genetics (Westborough, Massachusetts) also is developing a new test for assessment of cancer drug response. Genzyme Genetics is one of the leading providers of cancer testing services, having performed more than 175,000 tests on breast cancer patients to determine HER2 expression. The company performs about 1 million genetic tests per year in its reference laboratories, divided about equally between cancer and genetic disease testing. A new test for mutated EGFR is in development that will have important applications in guiding targeted drug therapy. While EGFR expression has not proven to be a highly reliable indicator of response to drugs that target that receptor, the presence of EGFR mutations has been shown to correlate quite closely with response in recent studies.

Early cancer detection technologies advance

In addition to guidance of therapy, analysis of protein expression patterns is showing promise as a method for early detection of cancer. Gil Mor, MD, of Yale University School of Medicine (New Haven, Connecticut) discussed a protein marker strategy for early detection of ovarian cancer, with a goal of developing an inexpensive screening test for the disease. Noting that many previous attempts at using array-based testing in cancer have suffered from lack of specificity of the predictive method, Mor described a multi-phase research program that began with a set of 169 candidate proteins identified by microarray analysis as potentially associated with ovarian cancer. That marker set was then narrowed down to 35 proteins that were known to be linked to ovarian cancer based on prior studies. The reduced set of markers was then validated on a separate group of patients, and only 10 were found that provided predictive power.

The six markers that showed the highest predictive power via microarray analysis were then re-evaluated using ELISA testing, leaving only four markers that correlated based on both ELISA and microarray assays. While none of the markers by itself has high sensitivity and specificity, the panel of four markers provides sensitivity of 96% in Stage I/II ovarian cancer and 95% in Stage III-IV. The key markers include prolactin, IGF-2, and leptin. In normal patients without cancer, 94% tested negative, and the four who tested positive were all at high risk for ovarian cancer. Mor noted that it would be necessary to further improve the sensitivity to 99.6% in order for the test to be useful in screening of the general population. The researchers are planning to add three proteins to the four-marker panel to improve sensitivity, and are also planning to develop an inexpensive point-of-care screening test.

Douglas Ross, PhD, of Applied Genomics (Huntsville, Alabama), described a panel of markers identified via gene expression profiling that could have utility in early detection of lung cancer. Using a similar strategy as that employed by Mor, the discovery program started with the evaluation of over 700 novel antibodies for potential lung cancer protein antigens, and narrowed that down to a 45-marker panel of diversity that was then evaluated for predictive power in two separate groups of patients. Ultimately, four antibodies were identified that could identify patients with squamous cell lung cancer, and an additional four were found to detect lung adenocarcinoma. Applied Genomics also is developing similar protein marker panels for detection of breast cancer.

A non-invasive approach for early detection of lung cancer was described by Michael Phillips of Menssana Research (Fort Lee, New Jersey) at the ASCO conference. The technology relies on identification of patients carrying specific genetic polymorphisms (CYP2E1 and CYP1A1) by analysis of volatile organic compounds in breath. A panel of 28 markers, most of which are alkanes, has been evaluated. Seventeen of the 28 markers showed reduced abundance in patients with lung cancer. Individuals to be screened breathe into a disposable mouthpiece for two minutes, and a control sample of room air is also collected. The test is most sensitive for patients with early stage lung cancer (Stage I and II). Thresholds are adjusted to give a high negative predictive value of 99.6%, resulting in a rather high false positive rate of 14%. However, the test may have value as an initial low-cost screen to identify those individuals who should be referred for a chest imaging procedure, while eliminating a high percentage of individuals with very low risk.

Another new technology for use in early detection of cancer was exhibited at the conference by DOBI Medical International (Mahwah, New Jersey). DOBI's ComfortScan System uses a new approach for the early detection of breast cancer that is intended for use as an adjunct to mammography in cases where mammography findings are suspicious for cancer but not definitive. The ComfortScan System uses infrared light from an LED source and a CCD camera detector to create an image using Dynamic Optical Breast Imaging (DOBI). The infrared image is a highly sensitive method for detection of angiogenesis, or the growth of new blood vessels, which occurs when a tumor begins to grow, creating a requirement for increased blood supply. The $140,000 ComfortScan is sold only in Europe, South America, Asia, and Canada at present, but is undergoing trials for regulatory approval in the U.S. The system applies low pressure to the breast, about 0.5 pounds or 10 mm Hg, vs. the 20 to 25 pounds applied during mammography, and detects nascent blood vessels having a diameter of 2 mm to 3 mm. The technology is intended to improve upon the sensitivity of mammography for early breast cancer detection, and is primarily applicable to screening of women between the ages of 18 and 50.

The use of genetic testing is playing an increasingly important role in cancer management, primarily to detect individuals at increased risk who will benefit from more intensive screening. At a session on hereditary cancer screening, speakers discussed the current status of genetic testing for colorectal, breast, and ovarian cancer, diseases for which effective tests have been introduced, including the BRCA1/2 tests for breast and ovarian cancer. As discussed by Harry Groen, MD, PhD, of University Hospital Groningen (Groningen, the Netherlands), genetic tests are now available for hereditary non-polyposis colorectal cancer (HNPCC), also known as Lynch syndrome, that allow the degree of risk to be determined based on the type of mutations present. The mutations also are related to the risk for endometrial cancer. Patients with a family history of Lynch Syndrome are tested via molecular analysis, and those who test positive can either be followed more closely with screening or can undergo prophylactic surgery.

In the Netherlands, a program has been established to offer testing to those individuals who, based on family history, have more than a 10% probability of harboring the mutations. About 60% of at-risk individuals are tested in that country, at a cost of EUR 620 each, since the cost of testing is covered by the national health insurance program. While uptake of such testing is lower in the U.S. at 43%, due to a lack of comprehensive insurance coverage for testing, a number of reference labs including Myriad Genetics (Salt Lake City) and DNADirect (San Francisco), now offer cancer-related genetic testing. As discussed by James Church, MD, of the Cleveland Clinic Foundation (Cleveland), at the ASCO symposium, many insurance carriers now provide partial reimbursement, paying about $700 for a $1,000 genetic test, with the individual paying the remaining $300 out of pocket. A high percentage of individuals in the U.S., however, are unaware of the availability of genetic cancer testing, and 45% of those interviewed in a survey described by Church stated that their primary care physician never has mentioned the availability of genetic screening. Consequently, there is a considerable educational effort needed in the U.S. in order to increase the utilization of genetic cancer screening.

Advances in automated cellular analysis technologies are proving to be important in the development of improved tools for early detection of disease recurrence in cancer patients. Veridex (Warren, New Jersey), a Johnson & Johnson (New Brunswick, New Jersey) company, and Immunicon (Huntingdon Valley, Pennsylvania) have partnered to commercialize the CellSearch system, which uses Immunicon's rare cell analysis technology to detect small numbers of circulating tumor cells. The CellSearch system analyzes of 7.5 ml whole blood sample for the presence of cells bearing specific surface markers, including EPCAM and CD45. Those markers uniquely identify epithelial cells, which are believed to represent cancer cells that have migrated into the bloodstream from a metastatic tumor.

Initial studies with the system demonstrated that detection of more than five circulating tumor cells was indicative of metastasis in patients with measurable metastatic breast cancer. The most recent studies have evaluated patients with non-measurable cancer, and also have shown that levels of more than five cells is predictive of metastasis as soon as one month after initiation of therapy. That represents a significant advance over existing methods such as radiographic imaging, since the latter technique may not indicate that metastasis is occurring until up to a year after treatment starts.

Another approach to rare cell analysis in cancer patients has been developed by researchers at Innsbruck University Medical School (Innsbruck, Austria) led by Stephen Braun, MD. Braun described the use of cytochemical analysis of tumor cells in the bone marrow of node-negative breast cancer patients, combining data from nine separate studies involving 4,703 patients over a period of 10 years. Analysis of a subgroup of patients who were not treated with adjuvant chemotherapy because of their node-negative status showed that those who were positive for disseminated tumor cells in bone marrow fared significantly worse than those who were negative. Mean survival was 35 months less for the cell-positive patients. The results provide Level 1 evidence that bone marrow micrometastases are a strong prognostic factor for poor outcome in node-negative breast cancer.

Researchers at the University of Maryland Cancer Center (Baltimore) conducted another study of circulating tumor cells in breast cancer, using a fluorescence microscopy assay of blood samples collected over a 24-month period following initial treatment. As reported by Katherine Tkaczuk, the assay can detect as few as one or two circulating cancer cells in a 15 ml to 20 ml blood sample, representing a 50-fold to 100-fold higher sensitivity than can be achieved with RT-PCR. In most cases the presence of circulating tumor cells correlated with tumor stage, with an 88% positive rate in Stage IV vs. 47% and 41% in Stage II and III respectively. Moreover, the number of circulating tumor cells correlated with response to therapy.

While most physicians believe much larger studies are needed to provide convincing evidence for routine use of circulating tumor cell assays in cancer management, there already is a significant level of utilization of the tests now offered by clinical reference labs. Quest Diagnostics (Teterboro, New Jersey) is now performing about 120 tests per month using the Veridex CellSearch system, at a charge of $550 to $580 per test. Veridex charges $175 for the reagents used to perform the CellSearch assay. Quest has an exclusive agreement with Veridex to serve as the reference lab provider of the CellSearch assay.

WaveSense (Irvine, California) is developing a circulating tumor cell analysis system based on the company's UltraPlatform, which uses the EpiSep family of devices, and in particular the EpiSep XRC device which captures cells using cancer cell surface-antigen specific antibody coated magnetic particles. Captured cells are then viewed via fluorescence microscopy. The company is focusing on development of methods that allow detailed characterization of circulating tumor cells. A related application is analysis of sentinel lymph node biopsy specimens, to detect small numbers of cancer cells indicative of metastasis.In that application, detection sensitivity is improved relative to conventional histological analysis of frozen sections to 99.9% vs.a 20% to 30% false negative rate for conventional methods.

Advances in radiation treatment for cancer

Another topic addressed at the ASCO conference is the use of various device-based therapies that can be used in combination with chemotherapy to improve the efficacy of cancer treatment. Combining radiotherapy with chemotherapy is believed to provide an enhancement of therapeutic effect over that achievable with either modality used alone. Combined radiotherapy and chemotherapy is now the standard of treatment, for example, for small cell lung cancer that cannot be surgically resected. For other diseases, such as non-small cell lung cancer, the use of adjunct radiotherapy is still controversial, but such treatment is generally indicated for patients with advanced Stage III or higher disease.

Combined chemotherapy and radiotherapy is also of value in cervix cancer, resulting in a 50% reduction in death rate. Concurrent chemotherapy/radiotherapy also is being used in head and neck cancer, as discussed by David Brizel, MD, of Duke University Medical Center (Durham, North Carolina). In addition to combining radiotherapy with conventional chemotherapy agents, studies are also evaluating the use of FGF-7 targeting as well as the use of EGFR and VEGF inhibitors in combination with radiotherapy. The market for radiotherapy equipment is continuing to exhibit strong growth, as shown in Table 5, as improvements in radiation delivery technology are making the modality safer, with reduced side effects and increased efficacy because of the ability to more precisely target radiation dose to the tumor.

The leader in the market, Varian Medical Systems (Palo Alto, California), has recently introduced systems that combine cone-beam CT imaging with a linear accelerator, allowing beam position to be adjusted based on soft-tissue landmarks. The cone-beam CT scans can be compared with reference CT scans to allow fine-tuning of the patient set-up. Varian has just begun installing the new system worldwide, with two combined radiotherapy/cone beam CT systems in Europe and five now in the U.S. The company expects to have 15-20 systems installed by the end of the year.

TomoTherapy (Madison, Wisconsin) also exhibited a combination CT/LINAC system at the ASCO conference. About 30 of the company's Hi-ART systems have been installed worldwide, at a price of somewhat over $3 million each. The system offers a high-speed multi-leaf collimator for delivery of intensity-modulated radiotherapy, with a 50 millisecond cycle time, and plans to develop a system that will allow a tumor to be followed in real time during an irradiation procedure to provide highly precise dose targeting. The system also can treat multi-focal tumors in a single procedure, vs. the multiple set-ups required with conventional LINAC systems.