Keywords
cervical consistency index - cervical length - spontaneous preterm birth - first trimester
ultrasound
Introduction
Spontaneous preterm birth (sPTB) is defined as delivery before 37 weeks of gestation.
It affects about 10% of all deliveries worldwide. Its consequences on neonatal morbidities
and mortality are observed in everyday obstetric practice.[1] Cervical length (CL) is a well-established predictor for preterm birth (PTB) for
about three decades.[2] The probability of sPTB in those with short cervices at mid-trimester scan is higher
in high-risk populations than in low-risk women.[3]
[4]
[5]
[6]
[7]
The low sensitivity of short CL at mid-trimester scan for sPTB in low-risk pregnant
females makes its value for screening in such populations controversial.[3]
[4]
[5]
[6]
[7] On the other hand, the biochemical marker of sPTB (like fetal fibronectin) has low
availability and high cost.[8] All of these points warrant further exploration in the pathophysiology of preterm
labor as we try to find a more reliable predictor.
It is well known that cervical softening precedes cervical shortening in both term
births and PTBs. Cervical softening involves a remodeling of the cervical microstructure,
characterized by changes in water content and the alignment of collagen within the
cervical stroma. Therefore, methods that detect cervical softening could serve as
better predictors for sPTB than CL measurement.[9]
[10]
[11]
The cervical consistency index (CCI) was first introduced by Parra-Saavedra et al
in 2011 as an objective way to assess cervical softness during pregnancy.[12] They measured the percentage of the anteroposterior diameter of the cervix before
and at maximum pressure applied with a transvaginal ultrasound (TVUS) probe. They
found that CCI is lower in women who deliver preterm compared with those who deliver
at term.
CL measurement as a predictor for PTB was first introduced in 1996 and became supported
by evidence 20 years later in 2015, after more than 400 publications.[2]
[13] On the other hand, CCI has been described for about 15 years, and to this day, fewer
than 20 articles have been published about it and it has not yet become established
for routine obstetric practice.
Most studies on CL and CCI have evaluated their roles during the mid-trimester scan
visit. Few studies have assessed their roles during the first trimester (T1) scan
visit.[14]
[15] The aim of our study was to evaluate the usefulness of CCI measurement in predicting
sPTB in low-risk pregnant women at their T1 scan. Additionally, the effectiveness
of CCI was compared with that of CL measured during the same visit.
The primary objective of our study was to evaluate the role of CCI during T1 (T1-CCI)
in predicting sPTB before 37 weeks and before 34 weeks of gestation in a low-risk
population. The secondary objective was to compare the effectiveness of T1-CCI with
that of T1-CL for sPTB prediction. We chose the T1 scan rather than the mid-trimester
scan because acceptance of TVUS scanning is higher in the former among our population.
To the best of our knowledge, our study is the first to prospectively evaluate the
ability of the T1-CCI to predict sPTB in a low-risk population.
Methods
Our study was a prospective cohort involving 518 low-risk pregnant women with singleton
pregnancies. It was conducted at a single tertiary center (Habashy 4D scan; Alexandria,
Egypt) from September 2022 to September 2024. We used three criteria for including
cases: (1) singleton pregnancy, (2) T1 (between 11 and 14 weeks of gestation), and
(3) low-risk pregnant women.
To ensure the enrolled cases are from a low-risk population, we applied these 11 exclusion
criteria: (1) history of PTB or spontaneous abortion at 16 weeks of gestation or later,
(2) cerclage in current or previous pregnancies, (3) maternal medical disorders, especially
preeclampsia, thrombophilias, systemic lupus erythematosus, and diabetes mellitus,
(4) fetal anomalies, (5) preterm premature rupture of membranes or oligohydramnios,
(6) placenta previa, (7) maternal Mullerian anomalies, (8) use of assisted conception,
(9) iatrogenic PTB in the current pregnancy, (10) CL of 20 mm or less at enrolment,
and (11) spontaneous abortion of the current pregnancy (delivery before 28 weeks of
gestation).
At the start of our study, 539 cases had been enrolled. Twenty-one cases were excluded
during follow-up; therefore, the total number of cases was 518. The reasons for exclusion
after enrolment included: development of preeclampsia (4 cases), development of gestational
diabetes mellitus (2 cases), development of preterm premature rupture of membranes
(4 cases), diagnosis of fetal gross anomaly (3 cases), undergoing cerclage (91 cases),
loss to follow-up (7 cases), having iatrogenic PTB (6 cases), and experiencing spontaneous
abortion (4 cases).
All participants provided informed consent before enrolling in our study. The study
received approval from the ethical committee for medical research at the Faculty of
Medicine, Alexandria University, Alexandria, Egypt. All participants were scanned
during the T1 (between 11 and 14 weeks of gestation) using a TVUS probe (RIC5–9A-RS,
General Electric GE; Voluson S10-Expert). The scans were performed by Ahmed Elhabashy,
an assistant professor of obstetrics and gynecology at Alexandria University and a
fellow of Medicina Fetal Barcelona, Spain, who has 15 years of experience in obstetric
and gynecological sonography.
CL was measured following this protocol[13]: (1) on the sagittal plane of the cervix, (2) using a zoomed image that fills about
two-thirds of the screen, (3) without applying probe or fundal pressure, and (4) measuring
from the internal os to the external os with trace calipers in millimeters.
The CCI was measured using the following protocol[12]: (1) follow the first three steps of the CL measurement protocol, (2) measure the
anteroposterior diameter of the cervix (in millimeters) perpendicular to its longitudinal
axis at the midpoint (this measurement is AP), (3) apply pressure with the probe to
the cervix until no more compression is possible, and hold this pressure for 10 seconds,
(4) repeat step 2 (i.e. measure the anteroposterior diameter of the cervix in millimeters
perpendicular to its longitudinal axis at the midpoint after the 10 seconds of pressure;
this measurement is AP'), and (5) calculate the CCI as the result of step 4 (i.e.
AP') divided by the result of step 2 (i.e. AP), multiplied by 100, and expressed as
a percentage. [Fig. 1] illustrates the steps for measuring CL and CCI.
Fig. 1 The diagram illustrates how to measure the cervical length and the cervical consistency
index.
We tracked our cases with phone calls every 2 weeks until delivery. Cases that were
lost to follow-up or met any exclusion criteria during their follow-up were excluded
from the study.
Statistical Analysis of the Data
The statistical analysis of the data was conducted using IBM SPSS software version
20.0 (IBM Corp, released 2011, Armonk, New York, United States). Categorical data
were summarized as counts and percentages. To compare the studied groups, the chi-square
test was used. For continuous data, normality was evaluated using the Kolmogorov–Smirnov
test. Quantitative data were presented as range, mean, standard deviation, median,
and interquartile range. For nonnormally distributed quantitative variables, the Kruskal–Wallis
test was employed to compare more than two groups, followed by post hoc analysis using
Dunn's multiple comparisons test for pairwise comparisons. A receiver operating characteristic
(ROC) curve was generated by plotting sensitivity (true positives, TP) on the Y-axis versus 1-specificity (false positives, FP) on the X-axis at various cutoff values. The area under the ROC curve reflects the diagnostic
performance of the test. An area greater than 50% indicates acceptable performance,
while an area near 100% signifies excellent performance. The ROC curve also enables
comparison of the performance between two tests. The significance level for all statistical
tests was set at 5%.
Results
[Table 1] presents the descriptive analysis of the studied cases across various parameters.
We classified the cases into three groups based on gestational age at delivery: full-term
(≥ 37 weeks), late preterm (34–37 weeks), and early preterm (< 34 weeks). A total
of 450 cases (87%) were full-term deliveries (≥ 37 weeks). Fifty-seven cases (11%)
were late preterm (≥ 34–< 37 weeks), and 11 cases (2%) were early preterm (< 34 weeks).
Table 1
Descriptive analysis of the cases studied based on various parameters in each group
according to GA at delivery
|
Total
(n = 518)
|
GA delivery
|
Test of significance
|
p
|
|
Full-term (≥ 37)
(n = 450) 87%
|
Preterm (< 34–< 37) (n = 68) 13%
|
|
Late (≥ 34–< 37)
(n = 57) 11%
|
Early (< 34)
(n = 11) 2%
|
|
Age (y)
|
20–41
|
22–39
|
21–41
|
20–40
|
H =
3.453
|
0.178
|
|
Min. – Max.
|
|
Mean ± SD
|
30.47 ± 5.79
|
30.30 ± 5.85
|
31.81 ± 5.17
|
30.91 ± 6.17
|
|
Median (IQR)
|
30 (26–36)
|
30 (25–35)
|
32 (28–36)
|
32 (27–35)
|
|
GA US
(wk, d)
|
11 wk 6 d–14 wk
|
11 wk 6 d–14 wk
|
12 wk–13 wk 6 d
|
12 wk 1 d–14 wk
|
H =
0.265
|
0.876
|
|
Min. – Max.
|
|
Mean ± SD
|
13 wk ± 5 d
|
13 wk ± 6 d
|
13 wk ± 4 d
|
13 wk 4 d ± 4 d
|
|
Median (IQR)
|
13.0
(12 wk 1 d–12 wk 5 d)
|
13.0
(12 wk 1 d–13 wk 5 d)
|
13.0
(12 wk 2 d–13 wk 4 d)
|
13 wk 6 d
(12 wk 6 d–13 wk 3 d)
|
|
Parity
|
173 (33.4%)
|
153 (34.0%)
|
17 (29.8%)
|
3 (27.3%)
|
χ
2 = 0.586
|
0.7446
|
|
P0 (nullipara)
|
|
P ≥ 1
|
345 (66.6%)
|
297 (66.0%)
|
40 (70.2%)
|
8 (72.7%)
|
|
BMI (kg/m2)
|
21–37
|
23–36
|
21–34
|
24–37
|
H =
0.200
|
0.905
|
|
Min. – Max.
|
|
Mean ± SD
|
29.77 ± 4.62
|
29.73 ± 4.76
|
30 ± 3.52
|
30.35 ± 4.20
|
|
Median (IQR)
|
29.70
(25.60–34)
|
29.6
(24.8–34.6)
|
30.40
(27.40–33.10)
|
29.70
(27.25–33.80)
|
|
GA delivery
(wk, d)
|
28 wk–40 wk 5 d
|
37 wk 1 d–40 wk 5 d
|
34 wk 1 d–36 wk 4 d
|
28 wk–33 wk 4 d
|
H =
178.599[b]
|
< 0.001[b]
|
|
Min. – Max.
|
|
Mean ± SD
|
38 wk 1 d ± 13 d
|
38 wk 5 d ± 9 d
|
35 wk 1 d[a] ± 5 d
|
31 wk 1 d[a] ± 13 d
|
|
Median (IQR)
|
38 wk 2 d
(37 wk 1 d–39 wk 4 d)
|
38 wk 4 d
(37 wk 3 d–39 wk 5 d)
|
35 wk
(34 wk 4 d–36 wk)
|
31 wk 2 d
(29 wk 5 d–32 wk 5 d)
|
|
T1-CL (mm)
|
23–42
|
24–40
|
23–42
|
24–39
|
H =
18.627[b]
|
< 0.001[b]
|
|
Min. – Max.
|
|
Mean ± SD
|
31.08 ± 5.29
|
31.42 ± 5.22
|
29.01[a] ± 5.42
|
27.82[a] ± 4.51
|
|
Median (IQR)
|
30 (27–35)
|
30.45 (27.10–35.40)
|
27 (24.40–32)
|
27 (24–29.5)
|
|
T1-CCI
|
0.66–0.96
|
0.71–0.96
|
0.69–0.81
|
0.66–0.83
|
H =
91.805[b]
|
< 0.001[b]
|
|
Min. – Max.
|
|
Mean ± SD
|
0.83 ± 0.09
|
0.84 ± 0.09
|
0.73[a] ± 0.04
|
0.70[a] ± 0.06
|
|
Median (IQR)
|
0.83 (0.73–0.92)
|
0.85 (0.76–0.93)
|
0.72 (0.69–0.76)
|
0.67 (0.66–0.73)
|
Abbreviations: BMI, body mass index; CCI, cervical consistency index; CL, cervical
length; GA delivery, gestational age at delivery; GA US, gestational age at the time
of ultrasound; IQR, interquartile range; SD, standard deviation; T1, first trimester;
χ
2, chi-square test.
Note: H: H for Kruskal–Wallis test, pairwise comparisons between each two groups were
done using the post hoc test (Dunn's for multiple comparisons).
a Statistically significant with full-term group.
b Statistically significant at p ≤ 0.05.
[Table 1] shows that there was no statistically significant difference between the three groups
based on maternal age, gestational age at the time of the ultrasound, parity, and
body mass index. T1-CL and T1-CCI were significantly lower among cases who delivered
preterm (68 cases) than those who delivered at full term (450 cases).
[Fig. 2] shows the ROC curves for T1-CL and T1-CCI. It illustrates the data presented in
[Table 2]. ROC curve [Fig. 2A] demonstrates the diagnostic performance of T1-CL and T1-CCI in differentiating preterm
cases from full-term cases. ROC curve [Fig. 2B] shows the diagnostic performance of T1-CL and T1-CCI in distinguishing early preterm
cases from late preterm cases.
Fig. 2 Receiver operating characteristic (ROC) curves illustrating the performance of first-trimester
cervical length (T1-CL) and first-trimester cervical consistency index (T1-CCI). (A) ROC curve for T1-CL and T1-CCI to distinguish preterm cases (n = 68) from full-term cases (n = 450). (B) ROC curve for T1-CL and T1-CCI to differentiate early preterm cases (n = 11) from late preterm cases (n = 57).
Table 2
Diagnostic performance of first trimester cervical length (T1-CL) and first trimester
cervical consistency index (T1-CCI) in differentiating preterm cases (n = 68) from full-term cases (n = 450), and early preterm cases (n = 11) from late preterm cases (n = 57)
|
|
AUC
|
p
|
95% CI
|
Cutoff
|
Sensitivity
|
Specificity
|
PPV
|
NPV
|
|
Discriminate all preterm cases (n: 68) from full-term cases (n: 450)
|
T1-CL
|
0.659
|
< 0.001[a]
|
0.583–0.735
|
≤ 24 mm
|
20.59
|
98.22
|
63.4
|
89.11
|
|
≤ 25 mm
|
32.35
|
83.78
|
23.16
|
89.13
|
|
≤ 26 mm
|
41.18
|
100.0
|
100.0
|
87.89
|
|
≤ 26.4 mm[b]
|
47.06
|
83.78
|
30.48
|
91.28
|
|
≤ 27 mm
|
51.47
|
75.33
|
23.97
|
91.13
|
|
T1-CCI
|
0.858
|
< 0.001[a]
|
0.817–0.900
|
≤ 0.73
|
66.18
|
76.89
|
30.20
|
93.77
|
|
≤ 0.74
|
67.65
|
76.89
|
30.67
|
94.02
|
|
≤ 0.75[b]
|
73.53
|
76.89
|
32.47
|
95.05
|
|
≤ 0.8
|
91.18
|
61.11
|
26.16
|
97.86
|
|
T1-CL (≤ 26.4 mm) and T1-CCI (≤ 0.75)
|
45.59
|
100.0
|
100.0
|
92.4
|
|
T1-CL (≤ 26.4 mm) and/or T1-CCI (≤ 0.75)
|
75.0
|
60.67
|
22.37
|
94.14
|
|
Discriminate late preterm cases (n: 57) from early preterm cases (n: 11)
|
T1-CL
|
0.554
|
0.571
|
0.384–0.724
|
≤ 28 mm
|
63.64
|
47.37
|
18.92
|
87.10
|
|
≤ 29 mm
|
72.73
|
40.35
|
19.05
|
88.46
|
|
≤ 30 mm
|
81.82
|
36.84
|
20.0
|
91.30
|
|
≤ 31 mm[b]
|
90.91
|
29.82
|
20.0
|
94.4
|
|
≤ 32 mm
|
90.91
|
24.52
|
18.87
|
93.33
|
|
T1-CCI
|
0.738
|
0.013[a]
|
0.521–0.954
|
≤ 0.66
|
45.45
|
100.0
|
100.0
|
90.48
|
|
≤ 0.67[b]
|
63.64
|
100.0
|
100.0
|
93.4
|
|
≤ 0.68
|
63.64
|
100.0
|
100.0
|
93.44
|
|
≤ 0.69
|
63.64
|
73.68
|
31.82
|
91.30
|
|
≤ 0.70
|
63.64
|
61.40
|
24.14
|
89.74
|
Abbreviations: AUC, area under the curve; CI, confidence interval; NPV, negative predictive
value; PPV, positive predictive value.
a Statistically significant at p ≤ 0.05.
b Cutoff was chosen according to Youden index.
[Fig. 2] shows that the diagnostic performance of T1-CCI surpasses that of T1-CL in distinguishing
cases who will deliver full term from those who will deliver preterm. The T1-CL cannot
differentiate between early and late preterm cases, but T1-CCI can distinguish between
them. [Fig. 2] provides a graphical representation of the data listed in [Table 2], which displays the areas under the curve (AUCs) of T1-CL and T1-CCI and their diagnostic
performances at different cutoffs.
The upper half of [Table 2] showed that the diagnostic performance of T1-CCI in distinguishing between those
who will deliver at full term and those who will deliver preterm is better than that
of T1-CL, as the AUCs were 0.858 (95% confidence interval [CI]: 0.817–0.90) and 0.659
(95% CI: 0.583–0.735), respectively. The optimal cutoffs based on the ROC curve and
Youden index for differentiating between full-term and preterm delivery were 26.4 mm
for T1-CL and 0.75 for T1-CCI.
At these cutoffs, the sensitivity of T1-CCI for predicting PTB was much higher than
that of T1-CL (73.53% vs. 47.06%), with nearly similar specificities (76.89% vs. 83.78%).
The specificity of combined T1-CL and T1-CCI is 100%, meaning that if T1-CL exceeds
26.4 mm and T1-CCI is greater than 0.75, the case is unlikely to deliver preterm.
The lower half of [Table 2] showed that T1-CL is not reliable for distinguishing cases that will deliver early
preterm from those that will deliver late preterm, as its AUC was 0.554 (95% CI: 0.384–0.724)
and the p-value was 0.571. Conversely, T1-CCI is reliable for making that distinction, with
an AUC of 0.738 (95% CI: 0.521–0.954) and a p-value of 0.013.
The optimal cutoff of the T1-CCI based on the ROC curve and Youden index to distinguish
between cases that will deliver early preterm from those that will deliver late preterm
was 0.67. At this cutoff, the sensitivity, specificity, positive predictive value
(PPV), and negative predictive value (NPV) of T1-CCI were 63.64, 100, 100, and 93.4%,
respectively. The cutoffs listed for T1-CL are clinically insignificant because the
AUC was low.
From [Table 2], we see that during the TVUS performed in the T1 scan, if the T1-CCI was 0.76 or
more, the case is most likely to deliver at or beyond 37 weeks; if the T1-CCI was
between 0.68 and 0.75, the case is most likely to deliver late preterm (34–37 weeks);
and if the T1-CCI was 0.67 or less, the case is most likely to deliver early preterm
(< 34 weeks). These cutoffs are shown in [Table 3].
[Fig. 3] shows examples of normal and abnormal T1-CCI. It depicts two different cases where
CCI was measured at the 14th week of gestation.
Fig. 3 Cervical consistency index (CCI) measured in the first trimester (T1) for two different
cases at 14 weeks of gestation. (A) Normal T1-CCI > 0.75. (B) Abnormal T1-CCI ≤ 0.75.
Table 3
Cutoffs of the first trimester cervical consistency index (T1-CCI) to predict gestational
age (GA) at delivery
|
Prediction of GA at delivery by first trimester cervical consistency index (T1-CCI)
using TVUS
|
|
T1-CCI
|
≥ 0.76
|
0.68 - 0.75
|
≤ 0.67
|
|
GA at delivery (wk)
|
≥ 37 wk
(full term)
|
≥ 34–< 37 wk
(late preterm)
|
< 34 wk
(early preterm)
|
Abbreviation: TVUS, transvaginal ultrasound.
Discussion
sPTB is a major cause of neonatal illness and death. The low sensitivity of short
CL for predicting sPTB in low-risk groups calls for more research into other predictors.
Since cervical softening occurs before its shortening, a marker for cervical flexibility,
such as the CCI, might help predict sPTB.
The aim of our study was to assess the role of CCI measurement in predicting sPTB
(< 37 and < 34 weeks) in low-risk pregnant women during their T1 scan. Additionally,
CCI's effectiveness was compared with that of CL measured during the same visit.
Our study was a prospective cohort that included 518 singleton pregnancies. CL and
CCI were measured using TVUS during the T1 scan (11–14 weeks of gestation). CL was
measured in the sagittal plane without applying probe pressure or fundal pressure,
from the internal to the external os, using trace calipers. The CCI is the ratio of
the anteroposterior diameter of the cervix at its midpoint, with maximal probe pressure
(maintained for 10 seconds), to the same diameter without probe pressure. Cases were
followed up by phone calls every 2 weeks until delivery.
T1-CL and T1-CCI were significantly lower in cases who delivered preterm (68 cases)
compared with those who delivered at full term (450 cases). The diagnostic performance
of T1-CCI was better than that of T1-CL in distinguishing cases who will deliver at
full term from those who will deliver preterm, as shown by the AUCs of 0.858 (95%
CI: 0.817–0.90) and 0.659 (95% CI: 0.583–0.735), respectively. The optimal cutoffs
based on the ROC curve and Youden index for differentiating between full-term and
preterm deliveries were 26.4 mm for T1-CL and 0.75 for T1-CCI.
At these thresholds, the sensitivity of T1-CCI for predicting PTB was significantly
higher than that of T1-CL (73.53% vs. 47.06%, respectively), with nearly comparable
specificities (76.89% vs. 83.78%). The combined specificity of T1-CL and T1-CCI is
100%, indicating that if T1-CL exceeds 26.4 mm and T1-CCI is greater than 0.75, the
case is unlikely to deliver preterm.
The T1-CL is unreliable for distinguishing cases likely to deliver early preterm from
those delivering late preterm, with an AUC of 0.554 (95% CI: 0.384–0.724) and a p-value of 0.571. In contrast, T1-CCI is reliable for differentiating between them,
showing an AUC of 0.738 (95% CI: 0.521–0.954) and a p-value of 0.013.
The optimal cutoff for the T1-CCI, determined by the ROC curve and Youden index to
distinguish between early preterm and late preterm cases, was 0.67. At this cutoff,
the sensitivity, specificity, PPV, and NPV of T1-CCI were 63.64, 100, 100, and 93.4%,
respectively.
From the data previously mentioned, we can see that during the T1 TVUS scan, if the
T1-CCI was 0.76 or more, the case will most likely deliver at 37 weeks or later; if
the T1-CCI was between 0.68 and 0.75, the case will mostly deliver late preterm (34–37
weeks); and if the T1-CCI was 0.67 or less, the case will mostly deliver early preterm
(< 34 weeks).
Becerra-Mojica et al[14] conducted a prospective study on the performance of T1-CCI to predict sPTB. To our
knowledge, this is the only published work on the CCI in the T1. Their study included
667 low-risk singleton pregnancies; 9.2% delivered before 37 weeks and 1.8% delivered
before 34 weeks. They found that, at a cutoff of 0.74, the sensitivity of T1-CCI to
predict sPTB before 37 weeks was 19.7%, and it was 33.3% for PTB before 34 weeks.
The specificity was 90.4% for sPTB before 37 weeks and 90% for sPTB before 34 weeks.
The NPV in their study was 91.8% for sPTB < 37 weeks, and it was 98.7% for sPTB < 34
weeks. The AUC in their study was 0.62 (95% CI: 0.54–0.69) for sPTB < 37 weeks, and
it was 0.8 (95% CI: 0.71–0.89) for sPTB < 34 weeks.
Our findings regarding the T1-CCI NPV and AUC were similar to what Becerra-Mojica
et al concluded in their study. However, there was a difference between our results
and theirs regarding the specificity and sensitivity of T1-CCI for predicting sPTB.
This discrepancy can be explained by two reasons. First, we added a step for CCI calculation
in our methodology: applying 10 seconds of probe pressure before measuring the AP
(the anteroposterior cervical diameter after probe pressure). Additionally, their
studied cases included both low-risk and high-risk populations, while our cases were
only low-risk.
Becerra-Mojica et al found that the T1-CL showed weak performance in distinguishing
cases who will deliver at full term from those who will have sPTB (p-value: 0.843). This finding aligns with our results.
Berghella et al[16] studied 183 high-risk singleton pregnancies to assess the performance of T1-CL in
predicting sPTB < 35 weeks. They used CL < 25 mm as the cutoff, similar to what we
used in our study. They concluded that the sensitivity, specificity, and NPV of T1-CL < 25 mm
were 14, 97, and 82%, respectively. Their results were similar to ours, despite studying
a high-risk population.
Antsaklis et al[17] studied 1113 low-risk singleton pregnancies to evaluate T1-CL for predicting sPTB.
They used CL < 27 mm as the cutoff. They concluded that the sensitivity, specificity,
and NPV of T1-CL < 27 mm for predicting sPTB before 37 weeks were 63, 51, and 91.6%,
respectively. The AUC was 0.6 (95% CI: 0.54–0.66). They also concluded that a short
cervix (T1-CL < 27 mm) did not have predictive value for PTB before 35 weeks (AUC = 0.55,
95% CI: 0.43–0.65).
Our findings regarding the T1-CL NPV and AUC were similar to what Antsaklis et al
concluded in their study. However, there was a discrepancy between our results and
theirs in terms of the specificity and sensitivity of T1-CL for predicting sPTB < 37
weeks and < 34 weeks. This discrepancy could be explained by four main factors. First,
differences in sample size. Second, we did not use fundal pressure during T1-CL measurement.
Third, we used a trace caliper for T1-CL, whereas they used a single-line measurement
method. Lastly, we used a 25-mm cutoff for T1-CL, while they used 27 mm.
We have five key strengths in our study. First, to the best of our knowledge, our
study is the first to prospectively evaluate the ability of the T1-CCI to predict
sPTB in a low-risk population. Our study was designed to screen only low-risk individuals,
where it remains debatable whether TVUS CL screening effectively predicts sPTB.
Our study is the first to establish cutoffs of T1-CCI for distinguishing between full-term
deliveries and sPTB, as well as between early and late preterm deliveries (0.75 and
0.76, respectively).
Another strength of our study is that we found the sensitivity of the T1-CCI is higher
than that of the T1-CL for predicting sPTB (73.53 and 47.06%, respectively). This
finding makes this marker (T1-CCI) an important predictor for sPTB. This high sensitivity
rate will result in a low FP rate and thus help prevent unnecessary interventions
for those with a short CL.
Additionally, the combined use of T1-CL and T1-CCI achieved 100% specificity in our
cases. The final strength of our study is that adding a 10-second probe pressure step
before measuring P' (P') significantly enhanced the sensitivity of T1-CCI for predicting
sPTB compared with previously published literature on the CCI methodology.
The main limitation of our study is the small sample size, which resulted in a low
number of preterm deliveries (68 cases). Another limitation is that all cases were
scanned by the same sonographer; therefore, more research is needed to assess the
reproducibility of CCI measurement among practitioners with different experience levels
and to determine interobserver variability. Larger multicenter studies are also necessary
before adopting T1-CCI as a prognostic indicator for sPTB in clinical practice.
Conclusion
T1-CCI is a better predictor for sPTB < 37 and < 34 weeks than the T1-CL in a low-risk
population. The optimal cutoffs for T1-CCI are 75% (for full term vs. preterm) and
67% (for early vs. late preterm). The high sensitivity of T1-CCI reduces the FP rate,
thereby avoiding unnecessary interventions. Further studies are needed before it can
be implemented in routine obstetric practice.
