Out of 1,000 women screened with 3D digital breast tomosynthesis (DBT) in US practice (BCSC 2023), about 83 are recalled (BI-RADS 0, vs. 103 to 116 with 2D digital mammography and 100 in the National Mammography Database). About 82% of recalls end without a biopsy recommendation - roughly 71% resolve right away as benign (BI-RADS 1/2) and 11.4% are placed in BI-RADS 3 surveillance (which carries a 1.86% 2-year cancer yield) - while 17.7% (~16 to 20 per 1,000 screened) are recommended for biopsy and 4.2 to 5.8 per 1,000 have a screen-detected cancer. For extremely dense breasts (Category D), full-protocol MRI halved interval cancers in the DENSE trial (NEJM 2019), and in the randomized BRAID trial (Lancet 2025), contrast-enhanced mammography (19.2 per 1,000; 15.7 invasive) and abbreviated MRI (17.4 per 1,000; 15.0 invasive) detected roughly 4 times as many cancers as automated ultrasound (4.2 per 1,000).
TL;DR: Preventive imaging works best when you know the base rates before the portal alert arrives.
- How 1,000 mammograms play out: Out of 1,000 women screened with 3D mammography (tomosynthesis), about 83 get called back for extra views (compared with 103 to 116 on 2D mammograms). About 82% of callbacks end without a biopsy - roughly 71% clear right away as normal or benign (BI-RADS 1 or 2) and 11% enter 6-month watch (BI-RADS 3, which carries a 1.9% two-year cancer rate) - while 16 to 20 women have a needle biopsy and 4 to 6 cancers are found.
- Dense breasts (Categories C and D) hide tumors on standard mammograms: Mammogram sensitivity drops from 90% to 93% in fatty breasts down to 78% to 82% in heterogeneously dense breasts (Category C) and 62% or lower in extremely dense breasts (Category D).
- Contrast MRI and contrast mammography catch 4 times as many cancers as ultrasound: In extremely dense breasts (Category D), full-protocol MRI cut interval cancers in half, and in head-to-head randomized testing, abbreviated MRI (17.4 cancers per 1,000) and contrast-enhanced mammography (19.2 per 1,000) caught four times as many cancers as automated ultrasound (4.2 per 1,000) at half the tumor size (10 mm vs. 22 mm).
- Pair breast imaging with a DEXA bone scan and heart screening: Bone loss speeds up at menopause, and heart disease remains the top mortality risk in women.
Almost every week in our clinic, a patient sends me one of two portal alerts that arrived late on a Friday afternoon. One is a screening mammogram callback stamped BI-RADS 0: Incomplete - Additional Imaging Needed. The other is a normal mammogram paired with the federal FDA MQSA notice saying she has dense breast tissue (Category C or D). Both alerts cause dread because the form letter never shares the base rates. About 82% of screening callbacks end without a biopsy (roughly 71% benign BI-RADS 1/2 and 11% BI-RADS 3 watch). And for dense breasts, randomized trials show how breast MRI and contrast mammography compare with ultrasound. Walking through those numbers together turns a scary alert into a clear plan.

What happens to 1,000 women screened with 2D vs. 3D mammography (BCSC & NMD Base Rates)?
Out of 1,000 women screened, about 83 are recalled with 3D tomosynthesis (DBT) versus 103 to 116 with 2D digital mammography (100 in the National Mammography Database); roughly 82% of recalls end without a biopsy (~71% BI-RADS 1/2 and 11.4% BI-RADS 3), 16 to 20 have a biopsy, 4.2 to 5.8 cancers are found, and 0.8 interval cancers are missed (Lehman 2017; Lee BCSC 2023; Lee NMD 2024).
The BCSC & National Mammography Database (NMD) Cascade per 1,000 Women Screened
US benchmarks come from three large registries: the 2017 BCSC 2D report (1.68 million screens, Lehman et al., Radiology 2017), the 2023 BCSC 3D tomosynthesis (DBT) vs. 2D report (Lee et al., Radiology 2023), and the National Mammography Database (NMD, 29.5 million screens, Lee et al., 2024; Berg et al., 2020):
| Step in the Screening Cascade (per 1,000 Screened) | 2D Digital Mammography (BCSC 2007-13 / 2011-18) | 3D Tomosynthesis (DBT, BCSC 2011-18) | National Mammography Database (NMD, 29.5M Screens) | What This Means for You |
|---|---|---|---|---|
| 1. Recalled for Diagnostic Workup (BI-RADS 0) | ~116 (11.6% in 2007-13) / ~103 (10.3% in 2011-18) | ~83 (8.3%) | ~100 (10.0%) | 3D tomosynthesis cuts callbacks by ~20 to 33 per 1,000 women vs. 2D mammograms. |
| 2. Resolved Right Away as Benign (BI-RADS 1 or 2) | ~82 (~71% of recalls) | ~59 (~71% of recalls) | ~71 (~71% of recalls) | About 7 in 10 recalled women clear right away as normal or benign (overlapping tissue or a simple cyst). |
| 3. Short-Interval Watch (BI-RADS 3) | ~13 (11.4% of recalls) | ~9 (11.4% of recalls) | ~11.4 (11.4% of recalls, Berg 2020) | Key nuance: Combining BI-RADS 1/2 (~71%) and BI-RADS 3 (11.4%) means ~82% of recalls end without a biopsy, though BI-RADS 3 carries a 1.86% 2-year cancer rate on 6, 12, and 24-month follow-up (Berg 2020). |
| 4. Recommended for Biopsy (BI-RADS 4 or 5) | ~20 (PPV2 = 25.6%) | ~18 | ~18 (17.7% of 2.87M recalls, Lee 2024) | Fewer than 2% of all screened women (17.7% of recalls) need a needle biopsy. |
| 5. Biopsy Performed (Core or Vacuum-Assisted) | ~18 (~13 benign biopsies) | ~16 (~10-11 benign biopsies) | N/A | PPV3 (cancers per biopsy done) is 35.5% with DBT vs. 31.7% (2011-18) / 28.6% (2007-13) with 2D. |
| 6. Breast Cancers Found (Invasive + DCIS) | 5.1 (2007-13) / 5.3 (2011-18) | 5.8 per 1,000 | 4.18 (~4.2) per 1,000 | PPV1 (cancer per recall): 6.9% with DBT vs. 5.2% (2011-18) / 4.4% (2007-13) with 2D. |
| 7. Cancers Missed (False Negatives Within 1 Year) | 0.8 per 1,000 (Sensitivity 86.9%, Specificity 88.9%) | 0.8 per 1,000 | N/A | False-negative misses cluster in dense breasts and fast-growing interval tumors. |
Three Key Realities About 10-Year Callbacks and BI-RADS 3 to 5
- The 10-Year Cumulative Callback vs. Biopsy Chance (Ho et al., JAMA Netw Open 2022): Over 10 years of BCSC screening, the chance of at least one false-positive recall is 49.6% with yearly DBT vs. 56.3% with yearly 2D mammography. With every-2-year screening, it drops to 35.7% (DBT) vs. 38.1% (2D). However, the 10-year chance of a false-positive biopsy recommendation is almost the same across modalities (11.2% yearly DBT vs. 11.7% yearly 2D). Screening interval, age, and breast density change your false-positive risk much more than 3D vs. 2D does. In extremely dense breasts, 10-year false-positive recall rates reach 65% to 67% with either test.
- BI-RADS 3 ("Probably Benign", <= 2% Risk) Has Strict Limits (Sickles 1991; Berg 2020; Lee 2021; Buch 2015): BI-RADS 3 applies only to validated nonpalpable findings after a full diagnostic workup (a smooth noncalcified mass, focal asymmetry, or tiny grouped calcifications on mammogram; an oval smooth parallel mass or complicated cyst on ultrasound; NCCN v1.2026). Among 45,202 women assigned BI-RADS 3 after recall in the NMD (Berg et al., Radiology 2020), the 2-year cancer rate was 1.86%. Fully 57.8% of cancers were caught at the 6-month visit, yet only 72.5% of women came back at 6 months. Because 88% of BI-RADS 3 cancers appear within 12 months (Buch 2015), standard follow-up is 6, 12, and 24 months (on screening ultrasound in ACRIN 6666, BI-RADS 3 cancer risk was 0.8%, Barr 2013). Age caveat: In the NMD (Lee et al., Radiology 2021), BI-RADS 3 cancer yield passes 2% starting at age 60 (and around age 54 for new findings), reaching 4.63% at ages 80 to 89.
- BI-RADS 4A-4C & 5 Biopsy Method and Concordance (Sharma Eur Radiol 2026; Liberman 2000; NCCN v1.2026): Cancer likelihood runs > 2% to <= 10% for 4A, > 10% to <= 50% for 4B, > 50% to < 95% for 4C, and >= 95% for BI-RADS 5. Ultrasound-visible masses get a 14-gauge ultrasound-guided core biopsy. Calcifications and distortions get a 9-to-11-gauge stereotactic or DBT-guided vacuum-assisted biopsy (VAB). VAB cuts underestimation of atypical ductal hyperplasia (ADH RR 0.63) and halves underestimation of DCIS (RR 0.47) compared with standard core biopsy (Sharma et al., Eur Radiol 2026). Every biopsy requires radiologic-pathologic concordance: 2% to 5% of needle biopsies are discordant, and ~24% of discordant benign biopsies prove malignant on repeat sampling (Liberman et al., Cancer 2000).
How does breast density (BI-RADS A to D) affect mammogram accuracy and supplemental screening?
Since September 10, 2024, the FDA MQSA rule requires every US mammogram report to tell you whether your breasts are not dense (Categories A and B) or dense (Categories C and D); mammogram sensitivity drops from 90%-93% in fatty breasts to 78%-82% in Category C and <= 62% in Category D (Holt JAMA 2026), where contrast MRI and contrast mammography find 4 times as many cancers as ultrasound.
What Randomized & Prospective Trials Show for Dense Breasts (DENSE, DENSE-2, BRAID, EA1141, J-START)
A key detail in the trial data: the Dutch DENSE trial tested full-protocol MRI. Abbreviated MRI was tested in a later multireader study of DENSE scans showing 84.3% vs. 85.9% sensitivity (van Grinsven Radiology 2025), and the ongoing DENSE-2 trial (van Grinsven Radiology 2026) randomizes women with Category D breasts to contrast-enhanced mammography (CEM) or abbreviated MRI. The head-to-head randomized trial comparing abbreviated MRI, CEM, and automated ultrasound is the UK BRAID trial (Gilbert et al., Lancet 2025):
| Clinical Trial & Design | Group Studied | Scans Compared After Negative Mammogram | Added Cancers Found (CDR per 1,000) & Interval Cancers | Recall, False Positives & Key Caveats |
|---|---|---|---|---|
| DENSE Trial (Bakker et al., NEJM 2019 Round 1; Veenhuizen, Radiology 2021 Round 2; RCT) | 40,373 Dutch women ages 50-75 with extremely dense breasts (Category D) and a negative 2D mammogram (1:4 randomization; 59% MRI uptake) | Full-Protocol Breast MRI (every 2 yr) vs. Mammography alone | Interval cancers halved: 2.5 vs. 5.0 per 1,000 by ITT (difference 2.5; USPSTF invasive interval cancer RR 0.47, 95% CI 0.29-0.77), and 0.8 per 1,000 (4 of 4,783) among women who completed MRI vs. 4.9/1,000 among decliners (complier causal reduction 4.2/1,000). Round 1 CDR: 16.5/1,000. Round 2 CDR: 5.8/1,000 (all Stage 0-I, node-negative). | Round 1: Recall 94.9/1,000, false-positive rate 79.8/1,000, biopsy 62.7/1,000 (PPV 26.3%). Round 2 false positives fell by two-thirds to 26.3/1,000. DCIS was ~10x more common in MRI arm (Longo NEJM 2019); 5 serious contrast/vasovagal events; €15,620/QALY q4y (Geuzinge 2021). |
| BRAID Trial (Gilbert et al., Lancet 2025; Interim RCT) | Women ages 50-70 with heterogeneously (C) or extremely (D) dense breasts and a negative screening mammogram | Abbreviated MRI (AB-MRI)** vs. Contrast-Enhanced Mammography (CEM) vs. **Automated Breast Ultrasound (ABUS) | Added CDR per 1,000: - AB-MRI: 17.4/1,000 (15.0 invasive, median size 10-11 mm) - CEM: 19.2/1,000 (15.7 invasive, median size 10-11 mm) - ABUS: 4.2/1,000 (all invasive, median size 22 mm, 0 DCIS) AB-MRI beat ABUS (p = 0.047) and matched CEM (p = 0.62). | Recall: 9.7% (AB-MRI) and 9.7% (CEM) vs. 4.0% (ABUS). CEM had 8.4 minor and 0.5 severe iodine contrast reactions per 1,000. Each extra cancer cost ~4-5 false positives with contrast imaging vs. ~8 with ABUS (Duffy 2025). |
| EA1141 (ECOG-ACRIN) (Comstock et al., JAMA 2020) | 1,444 women ages 40-75 with dense breasts (Category C or D) | Abbreviated MRI (< 10 min)** vs. **3D Tomosynthesis (DBT) | Invasive CDR: 11.8 per 1,000 (AB-MRI) vs. 4.8 per 1,000 (DBT). | Specificity: 86.7% (AB-MRI) vs. 97.4% (DBT). (In a paired reader study, Lawson Radiology 2023, CEM had lower recall than standard MRI [14.0% vs. 22.8%] and lower sensitivity [61% vs. 100%].) |
| J-START Trial (Ohuchi Lancet 2016; Harada-Shoji Lancet 2026; RCT) | 72,998 Japanese women ages 40-49 (median 11.4-year follow-up) | Mammography + Handheld Ultrasound vs. Mammography alone | CDR: 5.0 vs. 3.2 per 1,000. At 11.4 years, add-on ultrasound cut cumulative Stage II or higher breast cancer (dense subgroup interval cancers: 0.5 vs. 2.0 per 1,000; overall trial interval-cancer RR 0.58, 95% CI 0.31-1.08). | Specificity: 87.7% vs. 91.4%. In ACRIN 6666 (Berg 2012/2016), ultrasound added 5.3 cancers/1,000 in yr 1 (PPV3 < 10%) and 3.7 in yrs 2-3 vs. 14.7/1,000 for MRI; in ASTOUND-2 (Tagliafico 2018), ultrasound added 4.9 vs. 2.8/1,000 for DBT. |
How Do 2026 Guidelines Compare on Dense Breasts (Category C vs. Category D)?
Why do medical groups disagree? No supplemental-screening trial has shown a breast cancer death reduction yet. The proven endpoints so far are surrogates: fewer interval cancers (DENSE) and fewer Stage II+ cancers at 11.4 years (J-START). Also, a 2026 BCSC cohort (Kerlikowske et al., JAMA Netw Open 2026) found that community MRI raised early detection without cutting overall interval or advanced cancers. Decision modeling (Tosteson et al., Ann Intern Med 2026) showed that adding MRI to DBT for extremely dense breasts averts 0.1 to 0.8 deaths per 1,000 women at the cost of 22 to 186 extra false-positive biopsy recommendations, with the harm-benefit balance best when relative risk is >= 2:
| Guideline Body | Category C (Heterogeneously Dense, ~35-40% of Women) | Category D (Extremely Dense, < 10% of Women) | Key Nuance & Rationale |
|---|---|---|---|
| NCCN v1.2026 (BSCR-4) | Annual mammography with tomosynthesis; consider supplemental screening (Category 2A). | Annual mammography with tomosynthesis + Breast MRI with and without contrast starting at age 50 (can consider from 40); consider CEM or MBI if unable to have MRI; whole-breast ultrasound if CEM/MBI unavailable (Category 2A). | DENSE enrolled Category D only, while Category C lacks standalone RCT interval-cancer data. |
| ACR (2023 / Appropriateness Criteria 2024-2025) | Consider supplemental screening; Breast MRI preferred (CEM, MBI, or ultrasound as alternatives). | Supplemental Breast MRI preferred; CEM, MBI, or ultrasound as alternatives. | Prioritizes highest cancer sensitivity across C and D density. |
| EUSOBI (2022) | Shared decision-making. | Supplemental Breast MRI every 2 to 4 years for ages 50 to 70. | Every-2-to-4-year spacing cuts false positives and cost (€15,620/QALY at 4-yr interval). |
| USPSTF (2024) | I Statement (insufficient evidence for supplemental MRI or ultrasound). | I Statement (insufficient evidence). | Cites lack of mortality trials and risk of DCIS overdiagnosis / false-positive biopsies. |
| ACP (Version 2, 2026) | Recommends against routine supplemental MRI or ultrasound in average-risk women (consider DBT). | Recommends against routine supplemental MRI or ultrasound in average-risk women (consider DBT). | In average-risk women (relative risk < 2), false-positive biopsies (22-186/1,000, Tosteson 2026) outweigh modeled death reductions. |
Note: Why a palpable lump always needs a diagnostic mammogram + targeted ultrasound even after a normal screen: About 15% of breast cancers are interval cancers found between screens. Compared with screen-detected cancers, interval cancers are much more often Stage II or higher (43% vs. 12%), node-positive (22% vs. 12%), and triple-negative (21% vs. 6%) (Ambinder et al., Acad Radiol 2023). In women with a new lump within 6-12 months of a negative mammogram, targeted ultrasound finds a correlate in 50.3% vs. only 12.9% on repeat mammography (Leung AJR 2016), while diagnostic mammography still changes management in 35% of malignant lumps by mapping extent (Brown AJR 2017). See our guide on what to do when you find a breast lump.
Fishtown Medicine
A 90-minute conversation with Dr. Ash. A written plan you can follow.
What is the role of the DEXA scan and cardiovascular imaging in women?
While breast cancer screening gets the most attention, heart disease is the leading cause of death in women, and osteoporotic hip and spine fractures are a major cause of lost mobility after menopause.
- DEXA (Dual-Energy X-Ray Absorptiometry) Bone Density Scan: Measures bone mineral density (BMD) at the lower spine and hip (T-score >= -1.0 normal; -1.0 to -2.4 osteopenia; <= -2.5 osteoporosis) along with your 10-year FRAX fracture score. Standard guidelines start DEXA at age 65. Because women lose 10% to 20% of trabecular bone across the 5-to-7-year menopausal transition, we order an earlier baseline DEXA around perimenopause / early postmenopause (ages 50 to 55) or in any woman with low body weight, family hip fracture, early menopause, or steroid use.
- Cardiovascular Risk & CAC Scoring in Women: We pair cancer screening with ApoB, Lp(a), hs-CRP, and a coronary artery calcium (CAC) scan when risk is borderline or when female risk enhancers (early menopause < 40, history of preeclampsia or gestational diabetes) are present (see Heart Disease in Women).
Guidance from the Clinic
What are your actionable steps for preventive imaging?
Check your BI-RADS density letter (A to D), calculate your BCSC and Tyrer-Cuzick risk score, match supplemental MRI or CEM to Category D or high-risk Category C, and never let a recent normal screen delay a diagnostic workup for a new lump.
- Check your BI-RADS density letter (A, B, C, or D): Every US mammogram report since September 10, 2024 lists your density category under the FDA MQSA rule.
- Run your Tyrer-Cuzick and BCSC risk scores: If you have Category C or D density or any family history of breast or ovarian cancer, we calculate your 5-year, 10-year, and lifetime risk in clinic (Tosteson 2026 modeling shows supplemental MRI has the best benefit-to-harm ratio when relative risk is >= 2).
- Match add-on imaging to Category D vs. Category C (NCCN v1.2026): For extremely dense breasts (Category D), NCCN v1.2026 recommends Breast MRI with and without contrast starting at age 50 (consider from 40), or CEM / MBI if MRI-incompatible. For Category C, we decide on supplemental screening based on your combined risk score.
- Never wait on a lump you can feel: If you feel a new lump, even months after a normal 3D mammogram, book a diagnostic mammogram + targeted ultrasound (ultrasound finds a correlate in 50.3% of post-screen lumps vs. 12.9% on repeat mammography, Leung 2016).
- Anchor your baseline DEXA and lipid panel: Pair your mammogram schedule with a baseline menopausal DEXA scan and ApoB / Lp(a) check. Start your intake to build your schedule.
Key Takeaways
- The 1,000-Woman Mammogram Cascade (BCSC 2017/2023 & NMD 2024): ~83 recalled with 3D DBT (vs. 103-116 with 2D and 100 in NMD) -> ~82% end without a biopsy (~71% benign BI-RADS 1/2 + 11.4% BI-RADS 3 with a 1.86% 2-yr cancer yield) -> 17.7% (~16-20 per 1,000) recommended for biopsy -> 4.2 to 5.8 cancers found and 0.8 missed.
- BI-RADS 3 to 5 Rules (NCCN v1.2026; Lee 2021; Sharma 2026): BI-RADS 3 is for validated nonpalpable lesions after diagnostic workup (followed at 6, 12, and 24 months; risk exceeds 2% over age 60, reaching 4.63% at ages 80-89). BI-RADS 4A-5 goes to biopsy (vacuum-assisted biopsy cuts ADH upgrade RR 0.63 and DCIS upgrade RR 0.47) with mandatory concordance review (~24% of discordant benign biopsies are malignant).
- Extremely Dense Breasts (Category D, NCCN v1.2026 & DENSE): Full-protocol MRI halved interval cancers (2.5 vs. 5.0 per 1,000 ITT; 0.8/1,000 in MRI attendees, Bakker NEJM 2019). NCCN v1.2026 recommends Breast MRI starting at age 50 (consider from 40), with CEM/MBI second-line and ultrasound third-line.
- BRAID RCT & Risk Stratification (Lancet 2025; Tosteson 2026): Abbreviated MRI (17.4/1,000) and CEM (19.2/1,000) detect 4x more cancers than automated ultrasound (4.2/1,000) at half the size (10-11 mm vs. 22 mm), and the benefit-to-harm balance is strongest when combined relative risk is >= 2.
A note on cost: any discount we negotiate on labs and imaging passes straight through to you, with no markup. Our affordable labs and imaging guide covers how the billing works.
Scientific References
- BCSC & NMD Screening Benchmarks: Lehman CD, Arao RF, Sprague BL, et al. Radiology. 2017;283(1):49-58; Lee CI, Abraham L, Miglioretti DL, et al. Radiology. 2023;307(4):e222499; Lee CS, Goldman L, Grimm LJ, et al. (NMD). Breast Cancer Res Treat. 2024; Berg WA, Berg JM, Sickles EA, et al. Radiology. 2020;296(1):32-41; Lee CS, Berg JM, Berg WA. Radiology. 2021;299(3):550-558; and Ho TH, Bissell MCS, Kerlikowske K, et al. JAMA Netw Open. 2022;5(3):e222440.
- DENSE, DENSE-2 & Abbreviated MRI Trials: Bakker MF, de Lange SV, Pijnappel RM, et al. (DENSE Round 1). N Engl J Med. 2019;381(22):2091-2102; Veenhuizen SGA, et al. (DENSE Round 2). Radiology. 2021;299(2):278-286; van Grinsven SEL, Mann RM, Monninkhof EM, et al. Radiology. 2025 & Radiology. 2026 (DENSE-2); and Geuzinge HA, et al. JNCI. 2021.
- BRAID, EA1141, J-START & ACRIN 6666 Trials: Gilbert FJ, Payne NR, Allajbeu I, et al. (BRAID RCT). Lancet. 2025; Comstock CE, Gatsonis C, Newstead GM, et al. (EA1141). JAMA. 2020;323(8):746-756; Harada-Shoji N, Suzuki A, Ishida T, et al. (J-START 11.4-Yr). Lancet. 2026; Ohuchi N, et al. Lancet. 2016; and Berg WA, et al. (ACRIN 6666). JAMA. 2012 & JNCI. 2016.
- Biopsy Concordance, Community MRI & Modeling: Sharma N, Theis S, Vogelmann T, Pijnappel R. Eur Radiol. 2026; Liberman L, Drotman M, Morris EA, et al. Cancer. 2000; Kerlikowske K, Abraham L, Zhu W, et al. JAMA Netw Open. 2026; and Tosteson ANA, Stout NK, Su YR, et al. Ann Intern Med. 2026.
- Clinical Guidelines (NCCN v1.2026, USPSTF, ACR, EUSOBI, ACP): National Comprehensive Cancer Network. Breast Cancer Screening and Diagnosis (Version 1.2026); US Preventive Services Task Force. JAMA. 2024;331(22):1918-1930; Paulis LV, et al. (ACR). JACR. 2025; Mann RM, et al. (EUSOBI). Eur Radiol. 2022; and Qaseem A, et al. (ACP Version 2). Ann Intern Med. 2026.
Dr. Ash is a board-certified internal medicine physician specializing in preventive medicine and healthspan optimization at Fishtown Medicine in Philadelphia.
Frequently Asked Questions
Common Questions
Deep-Dive Questions
Ready when you are
The chat is our AI assistant, answering from our published guides. To talk it through with Dr. Ash himself, start with the intake.




