High circulating levels of IGF-1 are associated with an increased risk of ovarian cancer but only in premenopausal women, report Lukanova et al. (pages 549–554).
Ovarian cancer has the highest mortality rate of all gynecologic cancers, which is not only due to inadequate techniques for early detection and a lack of truly effective chemotherapeutic agents but also to a poor understanding of its etiology. The incidence of ovarian cancer appears to be weakly influenced by ovulatory history. Frequent ovulation is related to increased risk but additional hormonal factors such as steroid hormones, gonadotropins, insulin or insulin-like growth factor (IGF-1) also seem to play a role.
IGF-1 markedly stimulates the division and proliferation of normal and neoplastic cells and blocks the programmed self-destruction of cancer cells. Epidemiological studies related elevated circulating levels to an increased risk of breast, prostate and colon cancer. To test their hypothesis that IGF-1 levels may also be connected to ovarian cancer, the authors studied 132 women with primary invasive epithelial ovarian cancer from 3 prospective cohorts in the USA, Sweden and Italy.
Only women who had been diagnosed at least 1 year after blood donation and who did not use any exogenous hormones were included and matched with two control subjects. Lukanova et al. observed a strong association of IGF-1 levels with ovarian cancer risk among women who were younger than 55 at cancer diagnosis. They detected no association in the study population as a whole, or in women who were diagnosed with cancer after the age of 55.
These results mirror similar observations made in breast cancer studies, where elevated IGF-1 levels were linked with breast density and breast cancer risk exclusively in premenopausal women and women in whom cancer was diagnosed relatively early. The underlying reason for this age-effect still awaits detailed investigation.
EMERGING PATTERNS IN FAMILIAL BREAST CANCER
A new study confirms what had been suspected for some time: When stratified by etiology, breast cancer has a worse prognosis in BRCA1 mutation carriers. But those were also the patients who benefited most from oophorectomy, found Møller et al. (pages 555–559).
Almost 10 years ago, BRCA1 was the first identified gene to be linked to familial breast cancer. Since then it has been joined by at least 5 others but the prognosis of inherited breast cancer is still a matter of debate. Calculation of penetrance is complicated by the fact that an individual woman's risk is likely to be associated not only with the specific genetic mutation but also with many other risk factors, leading to exaggerated estimates.
Eleven collaborating European clinical centers dedicated to the early diagnosis and treatment of women at risk for inherited breast cancer joined forces to tease out the specifics of different underlying genetic factors and the effects of currently recommended interventions. Their latest report shows that BRCA1 mutation carriers present with biologically different tumors compared to other inherited breast cancers. Carriers have high-grade and estrogen-negative tumors, which are almost invariably of the infiltrating type.
Møller et al. conclude that genetic heterogeneity needs to be taken into account since individual groups of carriers may benefit differently from certain treatments and surveillance programs. The identification of additionally involved genes and of emerging patterns within carriers of different mutations will allow the development of custom-tailored treatments and will undoubtedly improve the outcome of familial breast cancers in the future.
A CLOSER LOOK AT SMOKING AND GASTRIC CANCER
After following their population-based cohort over 10 years, Sasazuki et al. could demonstrate that smoking is associated with an increased risk of the differentiated type of gastric cancer (pages 560–566).
Wide geographic variations indicate that the main risk factors for gastric cancer are environmental rather than of genetic origin. The recent worldwide decline in the number of gastric cancer cases can be entirely attributed to non-cardia tumors, supporting the hypothesis that cardia and non-cardia cancers are distinct entities with diverging etiologies. These discrepancies between subtypes of gastric cancer classified according to anatomic tumor site (cardia/non-cardia) and also to histology (intestinal/diffuse) emphasize the need to perform subtype-specific etiologic research.
Infection with Heliobacter pylori and a diet high in salt and low in fresh vegetables and fruit have been identified as major risk factors. A number of epidemiological studies pointed at smoking and alcohol consumption as additional contributing factors but most of them failed to differentiate between anatomic and histopathologic subtypes.
The present, prospective study did just that. It followed 19,657 Japanese men aged 40–59 years over a decade. Among them, 293 men were diagnosed with gastric cancer during this period. Past and current smokers appeared to have an elevated risk for differentiated types of distal gastric cancer compared to non-smokers, with adjusted rate ratios of 2.0 and 2.1, respectively. An increased risk for gastric cardia cancer of all histological types was suggested for alcohol consumption, but the association failed to reach statistical significance.
These results highlight the importance of studying large, prospective cohorts and of paying close attention to emerging patterns in subgroups stratified by relevant parameters.
THE RISE OF A NEW ANTAGONIST
The humanized anti-epidermal growth factor receptor (EGFR) antibody h-R3 is emerging as a promising new candidate from the growing pool of inhibitors that target oncogenic members of the epidermal growth factor receptor (EGFR) family. When Crombet-Ramos et al. tested h-R3 for its antiproliferative, antiangiogenic and proapoptotic activity in vitro and in vivo the antibody passed with flying colors (pages 567–575).
Ever since the observation was made that overexpression of both EGFR and its specific ligands leads to autocrine stimulation in a number of epithelial tumors, researchers have been looking for a means to interrupt this fateful feedback loop. The success of a humanized monoclonal antibody, known as herceptin/trastuzumab, against the EGFR family member HER-2/ErbB2 oncogene invigorated the search for inhibitors directed against EGFR itself. This effort led to the development of several EGFR antagonists (antibodies and small molecular weight compounds), among them Cetuximab/IMC-C225, Iressa/ZD139, PKI-166 and the humanized, neutralizing monoclonal antibody h-R3 investigated in the presented study.
When the authors applied h-R3 to monolayer cultures of the highly EGFR-dependent squamous carcinoma cell line A431, they detected a growth-inhibitory effect of 40%. But more importantly, h-R3 acted as a potent cytotoxic agent against A431 tumors in vivo. Mice carrying A431 xenografts were treated with 1mg h-R3, administered peritoneally every other day for 2 weeks. All animals showed complete remission by day 22 followed by an at least 2-week relapse-free interval. This effect was probably achieved through the observed 2-fold reduction of VEGF mRNA and protein levels, leading to a significant decrease in microvascular density, as well as the disappearance of dilated “mother vessels.”
Several ongoing clinical trials with h-R3 either as monotherapy or in combination with radiotherapy will show whether the promising candidate can hold its own outside the lab.
LET COMPUTERS MAKE THE DISCRIMINATION
Using a digital dermoscopy analyzer in combination with a trained artificial neuronal network to evaluate pigmented skin lesions, Rubegni et al. obtained a diagnostic accuracy of about 94% (pages 576–580).
In the last few decades, the incidence of cutaneous melanomas has increased significantly worldwide, particularly in the fair-skinned populations. Because of the high risk of metastasis in advanced cutaneous melanomas, early diagnosis of this neoplasm represents a continuous challenge for all physicians, especially dermatologists.
The introduction of dermoscopy or epiluminescence microscopy (ELM) surface microscopy, which allows the assessment of structures not visible under naked-eye examination, considerably increased the diagnostic accuracy for pigmented skin lesions (PSL). Nevertheless, the evaluation of the many morphological characteristics by PSL is complex and the correct interpretation of the ELM images still hinges on the experience of the viewer.
To circumvent the problem of subjectivity, Rubegni et al. trained an artificial neuronal network to correctly interpret 48 different parameters grouped into 4 categories (geometries, colors, textures and islands of color) obtained by a digital dermoscopy analyzer. A series of 371 benign and 217 malignant excised, clinically atypical, flat PSL was evaluated. The artificial neuronal network could distinguish between the two groups with an accuracy of 94% based on as few as 13 features.
This new approach is a step in the right direction: to enable the objective evaluation of the many parameters that have emerged after years of experience with dermoscopy.