If your first scan came after the first trimester, or if you are wondering what the later scans can and cannot do to your due date, the report looks different from an early one. There is no single crown-rump length, and the gestational age is built from several numbers.
This guide covers what those measurements are, how much error the sources attach to them, and why the rule for changing a due date gets more forgiving the later the scan is. It reports what the sources say and does not interpret any measurement. The first-trimester measurement has its own guide, crown-rump length dating explained.
What takes over from crown-rump length
StatPearls reports that crown-rump length (CRL) may be measured up to approximately 14 weeks, as the crown-rump length guide covers. After that, the report uses measurements of particular body parts instead. One of the original studies found that CRL after 14 weeks predicted age no better than these conventional measurements (Hadlock and colleagues, 1992).
StatPearls, an NIH reference, reports that second-trimester dating relies on a combination of measurements rather than any single one, following ACOG for the 14w0d to 21w6d range (StatPearls).
The four measurements on a report
The same StatPearls chapter describes how each is taken. On a report they usually appear as abbreviations:
| Label | Name | How StatPearls describes the measurement |
|---|---|---|
| BPD | Biparietal diameter | Across the head, from the outer to the inner edge of the skull |
| HC | Head circumference | Along the outer perimeter of the skull |
| AC | Abdominal circumference | Around the abdomen at the level where the portal vein divides |
| FL | Femur length | Along the thigh bone, perpendicular to the shaft, leaving out cartilage |
Hadlock's team developed equations that convert these measurements, alone or combined, into a menstrual age.
A 1984 study by Hadlock and colleagues of 361 fetuses between 14 and 42 weeks found that head circumference and femur length were the strongest single predictors of age, and that a number of combinations of measurements did significantly better than any single one. It also found that simply averaging the individual age estimates gave results not significantly different from a more complex equation. StatPearls adds that a 2024 validation study favored models using several measurements over those based on head circumference alone.
The margin of error, by stage
StatPearls reports ACOG's accuracy figures for ultrasound dating at each stage:
| Gestational age at the scan | Accuracy |
|---|---|
| 14w0d–21w6d | 7 to 10 days |
| 22w0d–27w6d | 10 to 14 days |
For comparison, StatPearls gives 5 to 7 days in the first trimester and 21 to 30 days from 28w0d, which it calls the least reliable method. The full comparison is in ultrasound vs. last period due dates.
The figures are margins of error, not corrections: the later the scan, the wider the band around the age the measurements give.
Why the margin for changing a due date widens
Under ACOG's approach, an established due date stays unless a scan disagrees with it by more than a set number of days. The Committee Opinion 700 abstract states the principle: once the date has been set from the last period, the first accurate scan or both, later changes should be reserved for rare circumstances, discussed with the patient and documented in the record.
An NIH trial protocol whose redating table follows ACOG's recommendations (CHAP protocol) lists the second-trimester rows for a first scan measured against the last-period date:
| Gestational age by last period at the first scan | Scan and last period agree if the gap is |
|---|---|
| 14w0d–15w6d | 7 days or fewer |
| 16w0d–21w6d | 10 days or fewer |
The protocol does not list rows past 21w6d. The full table and the rows for earlier scans are in ultrasound vs. last period due dates.
The threshold rises as the accuracy figure rises. StatPearls says the most accurate dating window is 8 to 12 weeks, described as the period of fastest and most consistent embryonic growth, and separately that the reliability of ultrasound dating declines as pregnancy progresses. The sources do not state the reason for the thresholds; the pattern is that the larger the scan's own margin, the larger the gap before it replaces the earlier date. For a first scan in the third trimester, StatPearls reports a gap of more than 21 days, together with the caution that a late size difference may reflect growth restriction rather than an error in earlier dating.
A worked example: the same gap at two scans
Say the first day of the last period was August 3, 2026. Adding 280 days gives a last-period due date of Monday, May 10, 2027.
- Scan at 15w0d. On November 16, 2026 the last-period count is 105 days, or 15w0d. The measurements give 16w1d (113 days). The scan's due date is November 16, 2026 plus 280 − 113 = 167 days, Sunday, May 2, 2027. The gap is 8 days against a threshold of more than 7 days (the 14w0d–15w6d row). The gap is outside the 7-day agreement range in the protocol's table, which is the case where ACOG's guidance treats the scan date as replacing the last-period date.
- Scan at 19w0d. On December 14, 2026 the last-period count is 133 days, or 19w0d. The measurements give 20w1d (141 days). The scan's due date is December 14, 2026 plus 139 days, also May 2, 2027. The gap is again 8 days, against a threshold of more than 10 days (the 16w0d–21w6d row). The gap is inside the threshold, so the last-period date stands.
The same 8 days lands on opposite sides of the line, because the threshold follows the scan's own margin of error. The ultrasound option in the site's due date calculator does the date arithmetic (scan date plus 280 minus the gestational age in days). Whether a due date changes is decided by your doctor or midwife, using their judgment on the whole picture.
What "suboptimally dated" means
ACOG calls a pregnancy suboptimally dated when no ultrasound confirms or revises the due date before 22w0d (Committee Opinion 700). The definition turns only on whether a confirming or revising ultrasound was done before 22w0d.
Committee Opinion 688 addresses such pregnancies. It states that the best clinical estimate of gestational age is the basis for decisions such as antenatal corticosteroid use and delivery timing, and that an interval ultrasound for fetal weight and gestational age 3 to 4 weeks after the first scan is reasonable to consider. Which steps apply is a question for your own prenatal care provider.
Bottom line
After about 14 weeks, ultrasound dating rests on four measurements (BPD, HC, AC and FL) combined into one estimate. ACOG's accuracy figures widen from 7 to 10 days through 21w6d to 10 to 14 days from 22w0d through 27w6d, and the gap needed to replace an established due date widens with them.
Find the gestational age on your own report and the date of the scan, then compare with the guide to anatomy scan timing to see where those dates fall in routine care. If your due date has moved or you are unsure which date your chart uses, ask your doctor or midwife.
Sources
- American College of Obstetricians and Gynecologists (via PubMed): Committee Opinion No 700: Methods for Estimating the Due Date (Obstetrics & Gynecology, 2017)
- American College of Obstetricians and Gynecologists (via PubMed): Committee Opinion No. 688: Management of Suboptimally Dated Pregnancies (Obstetrics & Gynecology, 2017)
- StatPearls, NCBI Bookshelf, U.S. National Library of Medicine: Sonography Assessment of Gestational Age
- PubMed, U.S. National Library of Medicine: Estimating fetal age: computer-assisted analysis of multiple fetal growth parameters (Hadlock et al., Radiology, 1984)
- PubMed, U.S. National Library of Medicine: Fetal crown-rump length: reevaluation of relation to menstrual age (5-18 weeks) with high-resolution real-time US (Hadlock et al., Radiology, 1992)
- NIH/NHLBI collaborative project, ClinicalTrials.gov NCT02299414: Chronic Hypertension and Pregnancy (CHAP) Clinical Protocol, v1.7 (Table 2, redating guidelines)
Newsletter
We ask only for your e-mail address. At most one e-mail a day, only when something new is published, with no tracking pixels. How we handle your address
