Published on

Published in: Health & Medicine
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide
  • I’d like to start with a 1-question survey. Is there anybody in this room who has not been in a car, either as a driver or a passenger, in the past month? If there is, please raise your hand. Of the 6000 or so people here I see ___ hands.
  • Let’s suppose that there is a medical condition for which everyone is at roughly equal risk worldwide and which each year in this country alone affects 5.3 million people and kills over 40,000. Now, let’s suppose there were an extremely effective, cheap, and non-toxic intervention to prevent this condition. It would be reasonable to assume that just about every one of us in this room would want to take advantage of this prevention opportunity. But we don’t. Not a single one of us does, as I’ve just shown. Because the condition I’m talking about is:
  • motor vehicle injuries. We could all get rid of our cars, arrange our lives so that we can walk to work, and use buses, trains and planes, which are much safer than cars, to get everywhere else. But we don’t. We drive and assume the risks - because we value the freedom and independence that our cars give us.
    Yet many well informed women at high risk for breast cancer, who decline prophylactic mastectomy because they value their breasts, are made to feel that their choice is irrational. It isn’t fair.
    We all do our best to avoid car accidents. Similarly, many high risk women use measures such as tamoxifen and oophorectomy to reduce their risk of breast cancer. But should breast cancer nevertheless develop, they hope that just as wearing a seatbelt significantly reduces the risk of dying in a car accident, screening will lower their risk of dying of breast cancer.
  • Hence the subtitle of this talk: Is MRI screening of the breast an effective seat belt for high risk women?
  • For the purpose of this talk I’ve defined ‘high risk’ as increased risk based on a known or likely inherited predisposition, because this is the only population in which MRI screening has been specifically studied. Whether these findings can be extrapolated to other high risk groups is a subject I’ll come back to at the end.
  • This graph, comparing the cumulative risk of breast cancer in subgroups of the high risk population and in the general population illustrates the two main challenges of screening very high risk women.
    The first challenge is the very young age at which we have to start screening. A 30 year old BRCA1 mutation carrier has about the same annual risk of breast cancer as a 60 year old woman in the general population.
    The 2nd challenge is the very high sensitivity required of the screening regimen. Of 100 woman in the general population getting screening mammography with 80% sensitivity, only 2 will have a breast cancer missed by screening before they reach age 70. But 13 of 100 BRCA1 carriers will have a cancer missed by screening, even if screening mammography were equally sensitive for them, and it isn’t.
  • Here are the current screening recommendations for women at high risk for hereditary breast cancer in the U.S. the U.K. and France. All 3 sets of guidelines are based on expert opinion only, because to date there is no evidence that any screening regimen reduces breast cancer mortality in the high risk population. Note that the only thing these guidelines have in common is a recommendation for annual mammography starting at an early age.
    NICE = National Institute for Clinical Excellence
    NCCN = National Comprehensive Cancer Network
  • What do we expect from mammography-based screening for the high risk population?
    Ideally it should have 100% sensitivity with no palpable interval cancers popping up between rounds of screening. The detected cancers should all be non-invasive, which means a cure rate approaches 100%. This is probably not realistic, especially for BRCA1 mutation carriers whose tumors tend to invade at a very early stage in their natural history, so we would want the invasive tumors to be no larger than 1 cm and always negative lymph nodes.
    Screening mammography doesn’t perform quite that well in the general population and, unfortunately in the high risk population it performs even less well. In 3 studies specifically focused on mutation carriers, mammography missed half the cancers, DCIS was rarely found, and a high percentage of the invasive cancers were large and node positive.
    There are two main reasons mammography is suboptimal in the high risk population.
  • The first reason is the very young age at which these women start screening.
    On average the younger a woman is, the greater the amount of radiodense fibroglandular tissue in her breast and the lower the sensitivity of mammography.
  • On the left is a mammogram of a fatty breast. Since fat is radiolucent it produces a dark background which makes the small, white stellate tumour in the center of the breast relatively easy to see. The mammogram on the right shows a very dense breast. It too has a cancer in roughly the same location but I challenge you to find it.
    To complicate the situation further, it appears that even among women of a given age, mammography is less sensitive in BRCA1 mutation carriers than in women with sporadic breast cancers.
  • Three separate studies of women presenting with palpable breast cancers have shown that the cancers of BRCA1 mutation carriers are half as likely to be visible on mammography as age-matched sporadic cancers.
  • This is probably due to the known biological differences between BRCA1 -related cancers, which often have a basal cell phenotype, and sporadic cancers, which usually have a luminal cell phenotype. BRCA1 cancers tend to have little or no associated DCIS so those tell-tale calcifications are absent on mammography. Also, BRCA1 cancers tend to be fleshy with round, pushing borders that give a more benign mammographic appearance than the typical spiculated borders of sporadic cancers.
  • For all these reasons, alternative breast screening modalities have been sought for high risk women and the most promising of these is MRI, which has several theoretical advantages over mammography.
    The intravenous contrast agent, Gadolinium DTPA, used with MRI concentrates in areas of tumor angiogenesis
    Breast MRI gives tomographic, 3_D images
    For both these reasons its sensitivity is much less affected by breast density.
    The lack of ionizing radiation is at the very least reassuring to many women.
    And they are delighted to learn that breast compression is not necessary.
  • But MRI also had disadvantages compared to mammography.
    The biggest one is its cost which I’ll come back to at the end.
    There are more false positives resulting in repeat imaging and unnecessary biopies.
    Biopsies under MRI guidance are technically more difficult than with mammography or ultrasound and may not be available at the facility that does the imaging. Some centres may do needle localization for surgical biopsies but not core biopsies.
    There are some nuisance logistics. To optimize sensitivity and specificity the study should be done in the 2nd week of the woman’s menstrual cycle.
    Very large women don’t fit into the magnet and there are the usual MRI restrictions on indwelling metal objects.
    About 5 to 10% of women get so claustrophobic that they require mild sedation and here’s why.
  • Thanks to some very encouraging early reports, in the mid to late 90’s at least 6 large studies were initiated in North America and Europe to look at the utility of breast MRI for screening women at high risk based on genetic status or family history.
  • These 6 studies are similar in several important ways which will allow future meta-analysis, and subgroup analyses. They are all non-randomized, not restricted to mutation carriers, and all include annual mammography and MRI done at most a few days apart.
    But these studies are also differ enough from each other for each to be interesting in their own right, specifically with respect to the number of centres involved, whether women with cancer are included, age range, whether ultrasound or clinical breast examination are included and the MRI technique.
    Interim results of two of these studies were published this year and I’d like to show these in some detail.
  • The Dutch National Study opened in Nov. ‘99 and involved 6 centres across the Netherlands. Women were eligible if they had never had cancer, were between the ages of 25 and 70, and had at least a 15% lifetime risk of breast cancers. Screening consisted of annual MRI and mammography plus 6-monthly CBE.
  • 19% were mutation carriers, of whom most had BRCA1 mutations Another 55% of the women were high risk and the rest were moderate risk
    ~ half the cancers were in mutation carriers and half the cancers were detected at the first round of screening
  • Opened Nov. ‘97
  • 73 % were mutation carriers
    37 cancers in 36 women
    32 of the 37 cancer were mutation carriers, evenly split between BRCA1 abd 2
  • In the Dutch study there was a non-significant trend towards lower sensitivity of MRI in the 2nd year of screening compared to the first, which we observed as well, but when subsequent years of screening are included it is clear that the superiority of MRI is persistent.
  • Since up to $50,000 per year of life saved is generally considered to be an acceptable cost for a medical intervention, even if the actual survival benefit is only half of my estimate, the cost is still reasonable.
  • Even if we show that MRI improves survival in the hereditary breast cancer population, there will be many remaining questions for further research.
    At what age should screening MRI begin and end? What is the optimal frequency? and should this differ according to mutation status or breast density?
    Air bags improve the effectiveness of seatbelts, but how much do mammography, ultrasound and CBE add to MRI?
    And finally, can results from the hereditary risk population automatically be extrapolated to other high risk populations? I would argue against such an approach and recommend that theses other groups be specifically studied. The biological differences between hereditary breast cancers and cancers arising in other high risk settings may affect the relative sensitivity of imaging modalities or the benefit of earlier detection.
    Some of these questions can probably be answered by meta-analysis of the results of the studies currently underway. The other questions will only be answered by future studies with these specific objectives.

    1. 1. Magnetic Resonance Imaging (MRI) Screening for High Risk Patients Ellen Warner M.D. Division of Medical Oncology Sunnybrook & Women’s College Health Sciences Center Toronto, Ontario, Canada
    2. 2. Each year in the U.S. alone: • 5.3 million affected • 40,000 deaths
    3. 3. Motor Vehicle Injuries Breast Cancer Primary Prevention: • obey traffic laws • tamoxifen • don’t drink & drive • oophorectomy Secondary Prevention: • seat belts ± air bags • breast screening
    4. 4. Is MRI Screening of the Breast an Effective Seat Belt For High Risk Women?
    5. 5. Definition of ‘High Risk’ • Known BRCA mutation carrier or • Close relative of mutation carrier or • Family history suggestive of inherited predisposition
    6. 6. Cumulative Risk of Breast Cancer 0% 10% 20% 30% 40% 50% 60% 70% 30 40 50 60 70 Age 1. Antoniou et al. Am J Hum Genet, 2003 2. SEER Cancer Stats Review, 2004. BRCA1 BRCA1 + oophorectomy general population no family mutation
    7. 7. High Risk Screening Guidelines U.S. (NCCN, 2004) U.K. (NICE, 2004) France (Eisinger, 2004) Mammography (annual) 25+ 30+ 30+ CBE (q 6months) 25+ - 20-25+ BSE (monthly) 18+ - - Ultrasound (annual) - - 30+ (dense breasts)
    8. 8. Mammography Screening for High Risk Women The Ideal • 100% sensitivity • DCIS • invasive ≤ 1cm, node -ve The Reality • 50% sensitivity • DCIS rarely found • 50% > 1 cm • 40% node +ve Brekelmans et al. JCO, 2001 Scheuer et al. JCO, 2002 Komenaka et al. Cancer, 2004
    9. 9. Limitations of Mammography for ‘High Risk’ Screening • young age = dense breasts
    10. 10. Mammographic Visibility of Palpable Breast Cancers 0% 20% 40% 60% 80% 100% BRCA1 sporadic Chang Lancet, ‘99 Goffin JNCI ‘01 Tilanus -Linthorst Int J Cancer ‘02 P=.03 P=.01 P=.01
    11. 11. Limitations of Mammography for HBC Surveillance • young age = dense breasts • tumour pathology (BRCA1) – less DCIS – fleshy, ‘pushing’ borders
    12. 12. Advantages of Breast MRI • Contrast agent concentrates in areas of tumor angiogenesis • tomographic images (3-D) • less influenced by breast density • no ionizing radiation
    13. 13. Disadvantages of MRI • $$$ • lower specificity • biopsy more difficult • logistics – menstrual phase – weight • claustrophobia
    14. 14. Breast MRI Screening Studies for High Risk Women Kriege et al. The Netherlands Kuhl, et al. Bonn, Germany Leach et al. U.K. Podo et al. Italy Schnall, Lehman et al. U.S. Warner, Plewes, et al. Toronto, Canada
    15. 15. Breast MRI Screening Studies for High Risk Women Similarities • prospective, non-randomized • not restricted to mutation carriers • annual mammography + MRI Differences • single / multiple centers • patient population • additional modalities • MRI technique
    16. 16. Dutch National Study Kriege et al. NEJM 351: 427, 2004. • 6 centers • unaffected women • ages 25-70 • ≥ 15% lifetime risk • MRI + mammography + CBE
    17. 17. Dutch National Study: Results • 1909 women – 358 mutation carriers – mean age 40 – mean # screens = 2 • 4 (9%) interval cancers! • 45 evaluable cancers • 39 invasive, 6 DCIS • 50% in carriers • 50% 1st screen
    18. 18. Sensitivity of Individual Modalities Dutch Study: Results 71% 40% 18% 0% 20% 40% 60% 80% MRI Mam CBE
    19. 19. Sensitivity: Invasive vs. In-Situ Dutch Study: Results 17% 80% 33% 83% 0% 20% 40% 60% 80% 100% Invasive In-Situ MRI Mammography n=39 n=6
    20. 20. False Positives Recalls Biopsies MRI 10% 5.8% Mammography 5% 1.7% Dutch Study: Results
    21. 21. Invasive Tumor Stage 40% 48% 43% 14% 12% 32% 37% 49% 25% 0% 20% 40% 60% 80% 100% Study Control 1 Control 2 > 2 cm 1.1 - 2 cm < 1 cm Dutch Study: Results n=45 n=1500 n=45 21% node + 52% node + 56% node +
    22. 22. Toronto Study Warner et al. JAMA 292: 1317, 2004 • single center • affected & unaffected women • ages 25 - 65 • >25% lifetime risk • MRI + mammography + CBE + US
    23. 23. The Toronto Study Medical Biophysics Donald Plewes PhD. Martin Yaffe PhD. Elizabeth Ramsay MSc Cameron Piron MSc Medical Imaging Petrina Causer M.D. Roberta Jong M.D. Belinda Curpen M.D. Joan Glazier MRT Garry Detzler MRT Caron Murray MRT Joanne Muldoon MRT Study Co-ordinator Kimberley Hill, BScGenetics Steven Narod M.D. Sandra Messner M.D. Wendy Meschino M.D. Andrea Eisen M.D. Pathology John Wong M.D. Judit Zubovits M.D. General Surgery Glen Taylor M.D. Claire Holloway M.D. Frances Wright M.D. Biostatistics Gerrit DeBoer PhD Alice Chung BSc Funding CBCRA NBCF Amersham Health Papoff Family Nurse Examiner Marg Cutrara R.N.
    24. 24. Toronto Study: Results • 437 women – 318 BRCA mutation carriers – mean age 43 – mean # screens = 3 Only 1 interval cancer! • 37 cancers – 32 in carriers – mean age 48 (34-64) – 28 invasive (2 lobular), 9 DCIS
    25. 25. Sensitivity of Individual Modalities Toronto Study: Results 84% 30% 8% 33% 0% 20% 40% 60% 80% 100% MRI Mam CBE US
    26. 26. Sensitivity of Combined Modalities Toronto Study: Results 97% 92% 92% 57% 38% 0% 20% 40% 60% 80% 100% All but CBE All but Mam All but US All but MRI Mam+ CBE
    27. 27. Sensitivity: Invasive vs. In-Situ Toronto Study: Results 86% 78% 25% 33% 50% 0% 0% 20% 40% 60% 80% 100% invasive In-Situ MRI MMG US n=28 n=9
    28. 28. Sensitivity by Age Toronto Study: Results 80% 88% 24% 35% 29% 45% 0% 20% 40% 60% 80% 100% <50 (n=20) 50+ (n=17) MRI MMG US
    29. 29. Toronto Study:: Results Sensitivity by Year of Screening 89% 79% 28% 32%28% 42% 0% 20% 40% 60% 80% 100% year 1 (n=18) year 2-5 (n=19) MRI MMG US
    30. 30. False Positives: Recalls Toronto Study: Results 1% 1% 19% 9% 2%2% 2% 6% 0.00% 5.00% 10.00% 15.00% 20.00% Year 1 Years 2 - 5 MRI M CBE US
    31. 31. False Positives: Biopsies Toronto Study: Results 11% 5%6% 6% 3% 16% 10% 4% 3% 1% 1% 0 0% 5% 10% 15% 20% Year 1 Year 2 Years 3-5 Any MRI M US
    32. 32. 74% 43% 22% 32% 3% 25% 0% 20% 40% 60% 80% 100% Toronto the Netherlands > 2 cm 1.1 - 2 cm < 1 cm Invasive Tumour Size
    33. 33. Toronto Study: Results Yr. # cancers DCIS Mean Invasive Size Node + 1 18 22% 1.1 (0.4 - 3.0) cm 3 2 9 11% 1.2 (0.4 - 2.0) cm 1 3-5 9 44% 0.8 (0.7 - 1.0) cm 0 Tumor Stage by Year No recurrences to date. Median f/u 3yrs. (range 1 to 7)
    34. 34. Effect of MRI Screening on Survival 0 5 10 15 1 2 3 4 5 years logcellnumber M e t s MRI mammo
    35. 35. Costs $$$ MRI Screening Benefits ???? Cost-Benefit Analysis
    36. 36. Cost-Benefit Estimate $$$ • 62 million women ages 30-60 in U.S. • 1% high risk (620,000) • $1200 per screen ____________________ $744 million/year • 620,000 high risk • 1% (6,200) have cancer • mortality 30% → 10% • 1240 more cured • mean years saved = 25 ________________________ 31,000 life years saved $24,000 / year of life saved
    37. 37. Summary Breast MRI for high risk women: • most sensitive screening modality • finds cancers at an earlier stage • has acceptable specificity • saves lives?
    38. 38. Other Research Questions • Optimal MRI screening schedule for subgroups? – age – mutation status – breast density • Role of other screening modalities? • Role of MRI for other high risk women? – Atypical hyperplasia, LCIS – Chest irradiation < age 30 – Very dense breasts