Breast Specific Gamma Imaging (BSGI) :A Novel Approach to Breast Cancer Detection

One in eight women will develop breast cancer

211,240 cases of IBC and 58,490 DCIS diagnosed this year

Rachel F. Brem, MD Director, Breast Imaging and Intervention Professor and Vice Chair Department of Radiology The George Washington University Washington, DC

96

Breast Cancer

Breast Cancer • 29% cancers in women

• 40,410 deaths annually

Five Year Survival for Breast Cancer Early Stage

95%

Late Stage

20%

• 18% of cancer deaths in women

Breast Cancer- Prevention • Drug therapy for prevention • Only true, currently available approach for prevention – All others are risk reduction

96% of breast cancers detected early can be treated successfully

Breast Cancer- Prevention

• Must detect when localized and curable

1

Mammograms lower mortality rate by up to 44%

Decrease in Mortality from Breast CA

The Most Widely Utilized Screening for Breast Cancer is Mammography • Breast cancer screening

• From 1975-1990 the death rate from breast cancer increased by 0.4% annually • Between 1990 and 2002 the rate decreased by 2.3% annually • Percentage of decline was greater among younger women – 1990-2002 decreased by 3.3% in women < 50 – 2.0% in women > 50 American Cancer Society Breast Cancer Fact and Figures 2005-2006

Decrease in Mortality from Breast CA • Meta-analysis on impact of improved screening and improved treatment (Chemotx) on decrease in mortality

–60% due to improved screening –40% due to improved treatment

Problems with Mammography • Do Not Diagnose Enough Breast Cancers – mammography is an IMPERFECT examination • 10-15% of breast cancers are not mammographically visible

Mammography • Decreased sensitivity (65%) in: – dense breasts – post surgical breast – implants

• 35% of breast cancer is NOT MAMMOGRAPHICALLY VISIBLE

Breast Cancer

How can we improve breast cancer detection

Problems with Mammography

Mammography

• Do Not Diagnose Enough Breast Cancers – mammography is an IMPERFECT examination • 10-15% of breast cancers are not mammographically visible

• only 20-30% of suspicious lesions prove to be malignant at biopsy

2

Breast Cancer

Breast Imaging

• How can we improve differentiating benign from malignant disease and thereby decrease the need for breast biopsy for benign lesions

Scintimammography (General Purpose Gamma Camera)

• Mammography – anatomic approach to imaging the breast – limitations

• Ultrasound – most common adjunct imaging modality – anatomic approach as well

Clinical

Objectives of the Two Scintimammography Breast Imaging Multicenter Clinical Trials

Scintimammography • nuclear medicine examination for the diagnosis of breast cancer • physiological vs anatomic examination – – – – – –

Miraluma (Bristol Myers Squibb) Uses traditional, multi-purpose gamma camera 99m technetium sestamibi FDA approved, 1997 identical to Cardiolite longstanding safety record

RESULTS Scintimammography Breast Imaging Trial:

100%

80%

•Determine the accuracy of Tc99m Sestamibi scintigraphic images in identifying malignant breast lesions

Reader 1 60%

Reader 2

40%

• Patients with palpable abnormalities

Reader 3

20%

• Patients with non-palpable, mammographic

0% SENSITIVITY

SPECIFICITY

PPV

NPV

ACCURACY

abnormalities

Clinical Miraluma™ Breast Imaging Trial: Dense vs Fatty Breast Tissue Non-Palpable Abnormality

Miraluma™ Breast Imaging Trial: Non-Palpable Abnormality

Scintimammography Clinical Imaging Trial

RESULTS

Breast Imaging

100%

SENSITIVITY 100%

100% 80%

80%

Fatty

Reader 1 60%

60%

Dense

Reader 2 40%

40%

Reader 3 20%

20% 0% READER 2 SENSITIVITY

READER 3

READER 1

READER 2 SPECIFICITY

READER 3

Reader 1

60%

Reader 2

40%

Reader 3

20%

0% READER 1

80%

0% SENSITIVITY

SPECIFICITY

PPV

NPV

ACCURACY

10 mm

85% (11/13)

92% (12/13)

Brem RF, Schoonjans JM, Kieper DA, Majewski S. et al High-resolution scintimammography: A pilot study. J Nucl Med 2002; 43:909-915.

Clinical Trial

Conclusions:

• Conclusions:

• Overall sensitivity of breast cancer detection improved

• Overcome intrinsic limitation of nuclear medicine imaging of the breast in clinical practice

• Improvement greatest in sub-centimeter cancers • HRBSG camera improved detection of < 1cm cancers to a level comparable to all cancers detected with the conventional gamma camera

What about screening high-risk women? • Improve sensitivity of mammography in women at increased risk – initially designed for dense breasts – at initiation of trial included all women at increased risk

• Adjunct to improve diagnosis of breast cancer in women at increased risk

• Sub-centimeter resolution • Imaging in mammographic position • Optimize breast imaging • Integration of Nuclear Medicine Imaging of the breast in clinical practice

High Resolution Nuclear Imaging of the Breast • Clinical Trial at GW Medical Center – High risk women (equal to STARR trial) – Normal mammogram and clinical examination – Annual “screening” with high resolution nuclear medicine camera • Abnormalities will be evaluated with ultrasound • Can be localized intra-operatively with pre-operative injection Brem RF, Rapelyea JA, Zisman G et al. Occult Breast Cancer: Scintimammography with High-Resolution Breast-specific Gamma Camera in Women at High Risk for Breast Cancer Radiology,2005,237

5

Materials and Methods

• Criteria for inclusion – normal mammogram (BI-RADS 1 or 2) – normal physical examination

Materials and Methods • 25-30 mCi Technitium sestamibi injected in the antecubital vein • Imaging immediately following injection for 6-8 minutes per image • Patient sitting

• True Negatives – All 78 patients – Normal mammogram, physical exam, and scintimammogram at year 1 and year 2 imaging

• False Negatives

94 Total Patients • 78 (83%) Negative BSGI • 16 (17%) Positive BSGI

Brem, Rapelyea, Zisman et al. Occult Breast Cancer: Scintimammography with High-Resolution Breast-specific Gamma Camera in Women at High Risk for Breast Cancer Radiology,2005,237

Brem RF, Rapelyea JA, Zisman G et al. Occult Breast Cancer: Scintimammography with High-Resolution Breast-specific Gamma Camera in Women at High Risk for Breast Cancer Radiology,2005,237

Results

Results

Results • False Positives – 14 (88%) of the 16 patients with positive BSGI determined to be benign • 5 patients confirmed by negative ultrasound • 9 patients confirmed by benign biopsy

– Normal scintimammogram with abnormal mammogram, ultrasound, or physical exam, who were found to have cancer at biopsy – 0 false-negatives

– All were followed for 1 year subsequent to BSGI, mammogram and PE

Fibrocystic Changes

Fibroadenoma

Results • Histopathology of biopsy proven falsepositives –7 with fibrocystic change • 1 with concomitant sclerosing adenosis

–1 fibroadenoma –1 fat necrosis

Fat Necrosis

6

Results

Results • True Positives –2 (13%) out of the 16 patients with positive BSGI –Histopathologically infiltrating and intraductal carcinoma

Results

• 2 Cancers –Detected only with BSGI –Both had a prior history of breast carcinoma

To determine the sensitivity and specificity of breast specific gamma imaging (BSGI) for the detection of breast cancer, using pathology as the reference standard

– 6mm lesion identified with ultrasound at location of focal uptake with scintimammogram – Pathologically measured 8 mm IDC

• One was a local recurrence • One was a contralateral cancer

Results • True Positive-Patient # 2 – 8 mm lesion identified with ultrasound at location of focal uptake with scintimammogram – Pathologically measured to be 12 x 10x 10mm

BSGI for Breast Cancer Detection

Results • True Positive-Patient # 1

Materials and Methods

BSGI Screening • 100% sensitivity • 84.8% specificity • 100% negative predictive value

Material and Methods

• Clinical indications for BSGI – palpable finding with no mammo correlate – evaluation of multicentricity/multi-focality in women with biopsy proven cancer – Equivocal mammographic finding – screening women at high risk for breast cancer

• Retrospective review of 146 women (age 32 to 98) undergoing BSGI • breast biopsy was performed

7

Material and Methods • Patients underwent BSGI with intravenous injection of 30mCi of 99mTc-sestamibi • Imaged in CC and MLO projections (7-10 min/image)

Materials and Methods • Studies were classified as: – positive (focal increased radiotracer uptake) – negative (no focal increased radiotracer uptake or scattered heterogeneous physiologic uptake)

• compared to biopsy results

Cancers

Sensitivity of BSGI

• 83 malignant lesions of which

• 83 malignant lesions (invasive carcinoma or DCIS) • BSGI identified 80 as malignant

– 67 (80.7%) invasive cancers – 16 (19.3%) DCIS

Sensitivity = 96.4% (95% CI, 89%-99%)

Invasive Cancer • • • •

Recently completed study 146 invasive cancer BSGI detected 143 Sensitivity = 98% – Only non-visualized cancers were Grade 1 and subcentimeter • 40% of sub-centimeter cancer, Grade 1 were visualized • N= 5, need larger study

BSGI Sensitivity of DCIS • 16 DCIS • BSGI identified cancer in 15

Sensitivity of 93.8 % (95% CI 69-99%)

Results • 146 patients • 167 lesions – 18 patients underwent biopsy of multiple lesions: • 1 patient with four biopsies • 1 patient with three biopsies • 16 patients with two biopsies

BSGI Sensitivity of Invasive Cancers

• 67 invasive cancers • BSGI identified cancer in 65

Sensitivity of 97.0 % (95% CI 89-99%)

BSGI and Cancer Size Of the cancers whose size was available • the mean size of invasive cancer detected by BSGI was 20 mm (n=56, SD 14 mm, median size 15mm) • Mean size of DCIS detected with BSGI was 18 mm (n=9, SD 18 mm, median size 7mm).

– Smallest cancer 2 mm

8

BSGI and Cancer Size Both the smallest invasive cancer and the smallest DCIS detected by BSGI was 1mm

Sensitivity of BSGI in Sub-Centimeter Cancers BSGI correctly identified: 16/18 cancers less than 1cm

Sensitivity = 88.9% *5 invasive cancers and 3 DCIS less than 5 mm were detected with BSGI

Lt CC

Lt MLO

•BSGI of 73 year old: Focal radiotracer uptake (circles) in lower outer left breast • Pathology demonstrated multifocal DCIS with no focus larger than 4mm.

Negative Predictive Value • Of 53 patients with a negative BSGI exam for malignancy

Rt MLO

Rt CC

Occult Cancers Detected only with BSGI • BSGI detected occult cancers not visualized with mammogram or ultrasound in 6 patients (7.2% of cancer patients) • In all 6 the lesion was found with second look ultrasound and underwent ultrasoundguided biopsy

Non-Malignant Lesions • 84 lesions

– 50 had no evidence of DCIS or invasive cancer,

– 82 normal or benign

NPV = 94.3% (95% CI, 84%-98%)

– 2 high risk • 1 ALH • 1 LCIS

46 year old woman: Focal increased radiotracer uptake (arrows) in the upper right breast Pathology demonstrated 0.6 cm infiltrating lobular carcinoma with extensive LCIS

Positive Predictive Value • Of 114 patients with a positive BSGI exam – 80 were invasive cancer or DCIS

PPV of 70.2% (95% CI, 60%-78%)

Specificity of BSGI 84 nonmalignant lesions • BSGI was negative in 50 • Positive in 34

Specificity of 59.5% (95% CI 48-70%)

– Both confirmed at surgery

9

False Positive Lesions 34 False Positive Lesions – Documented by biopsy – Most common pathology is Fibrocystic Change – 8 patients with FP had a biopsy in the preceding 2 months in area of increased radiotracer uptake • ? Inflammatory change

False Negative Studies 3 cancers • 1 DCIS, 2 IDC (3-10 mm) – 1 DCIS: High grade, measured 10 mm and was detected mammographically with retroareolar microcalcs – 2 IDC • 7 mm: Axillary tail (? positioning) • 3 mm: Incidental cancer found at prophylactic mastectomy in a patient with a contra-lateral breast cancer. – not identified with mammography, ultrasound or clinical examination

Conclusions

Conclusions: BSGI vs MRI

• 7.2% of patients with cancer had occult foci detected with BSGI not detected with other imaging modalities • Given the high sensitivity of BSGI, it can be considered as a pre-surgical exam in patients with biopsy proven cancer to look for additional foci as well as contralateral breast cancer.

• Our study supports the use of BSGI as MRI would be used in clinical practice with equal sensitivity and higher specificity • Greater ease for the patient • 4-8 images as compared to hundreds of images • Easily integrated into a breast imaging site

BSGI • Sensitivity = 97.0% for invasive cancers • Sensitivity 93.8% for the detection of DCIS • This sensitivity is comparable to that reported in MRI for invasive cancers (90.9%) and DCIS (93%) – Although larger study populations are needed, these findings support the potential of BSGI

DCIS: Mammo, MRI, BSGI and DCIS • 20 women • 22 biopsy proven DCIS –Size: 2-23 mm (mean 9.9mm)

DCIS: Sensitivity • Mammography • MRI

82% 88%

• BSGI

91%

Brem RF, Fishman MC, Rapelyea JA et al, Academic Radiology 2007; 14 :945-950

10

DCIS and BSGI • Smallest DCIS detected with BSGI: 2 mm • 2 occult DCIS lesions detected only with BSGI • 2 false negative DCIS lesions with BSGI – Both detected with microcalcifications mammographically

DCIS and MRI – 7 patients with 8 biopsy proven DCIS – 7 areas of abnormal enhancement (sensitivity 88%) • occult contralateral DCIS in one patient with bilateral disease – Detected initially with BSGI

• one false negative MRI examination, which at surgical excision demonstrated a 4 mm DCIS • false negative MRI was positive with BSGI.

BSGI vs MRI • 23 patients with 33 indeterminate lesions – Indeterminate breast finding requiring MRI and BSGI as part of their work up – 9 pathologically proven cancers in 8 patients Brem RF, Petrovitch I, Rapelyea JA. The Breast Journal, 2007 ;13 465-469.

BSGI vs MRI • • • •

4 ductal carcinoma in situ, 3 infiltrating ductal carcinomas 1 invasive lobular carcinoma 1 infiltrating carcinoma with duct and lobular features

BSGI and MRI

7 false positives BSGI lesions: • Sensitivity: Equal • Specificity: 27% MRI 75% BSGI

Infiltrating Lobular Carcinoma Right MLO Right CC

BSGI vs MRI – 6 FC – 1 sub-clinical abscess following TRAM reconstruction

18 false positive MRI studies lesions – – – – –

11 FCD 4 lobular neoplasia 1 fibrosis with foreign body giant cell reaction 1 abscess 1 no lesion visualized at biopsy with long term follow up

Infiltrating Lobular Carcinoma • 4 Institutions

•Difficult to identify mammographically •Difficult to palpate clinically •Lower sensitivity with MRI than other invasive cancers( 60% vs 90%)

– 2 academic, 2 private practice

• 26 women (ages 46 to 82 (mean age 62.8) with 28 biopsy proven pure ILC • mean size of 22.3mm (2mm-90mm)

Right Lateral

11

Infiltrating Lobular Carcinoma • Mammograms: Negative in 6/28 (21%) • Abnormal mammographic findings, 22/28 – 13 asymmetric densities – 4 architectural distortions – 5 spiculated masses

Infiltrating Lobular Carcinoma

Infiltrating Lobular Carcinoma

• Ultrasound (n=25) • 17/25 focal hypoechoic areas • Sensitivity = 68%

• MRI (n=12) • 10/12 lesions demonstrating enhancement • Sensitivity of MRI was 83%.

• Mammography had a sensitivity of 79%.

Known RT ILC No other known abnl

Invasive Lobular Carcinoma

Infiltrating Lobular Carcinoma

Rt. MLO

Right CC Right CC

Right MLO

• BSGI : Focal tracer uptake 26/28 cases • Sensitivity = 93%.

Left MLO Left CC Left CC Left MLO

Breast Specific Gamma Imaging • Clinical Uses: – All newly diagnosed breast cancer • Surgical planning • Occult foci of cancer

– High Risk Screening

BSGI

BSGI and Assessment Post Lumpectomy

• Assessment of positive margins following lumpectomy – Surgical planning – Extent of residual disease

• Even with normal mammo and/or PE

– Equivocal mammographic finding – Positive axillary adenopathy with no known primary

12

BSGI and Direct Silicone Injection

State of the Art • • • •

Commercially available FDA approved Reimbursed Numerous units are currently installed and more on order • It is HERE AND NOW!!!!!

• Extremely difficult mammographic interpretation • Adjunct imaging modality needed

Comparison to PET • Radiotracer cost – Have dose delivered every morning – Available and on hand

• Availability of radiotracer and dose • Minimal modifications to allow radiotracer in breast center (dose comparable to sentinel node) • Cost • Reimbursement

BGSI: Ongoing Studies

Localization of Area of Focal Uptake • Must localize or biopsy for integration into clinical practice – Optimally must be able to biopsy minimally invasively

• Technology for minimally invasive biopsy exists – translation of mammographically obtained stereotactic biopsy

BSGI:Conclusions • Molecular Imaging of the Breast • Multiple Clinical Indications • Important Adjunct Imaging Modality for the improved diagnosis of breast cancer • Easily and effectively integrated into Breast Center

Fusion Imaging

• Lymph Node Assessment – Perhaps decrease number of surgical procedures

• Improved differentiation of benign from malignant – Background to lesion ratio

• Response to Neo-adjuvant chemotherapy • Fusion imaging Courtesy of Jefferson Lab, Hampton University and Riverside Regional Medical Center

13