Keywords: obstetrics, cervix, hormones, t-test

Almost one in eight - babies born in the Azerbaijan are preterm, leading to complications for the mother and the baby, both during birth and development. Currently, there is no way to predict preterm labor, making its prevention and treatment virtually impossible.

Preterm birth is the second leading cause of infant mortality, causing over 17% of all infant deaths. It is associated with over 75% of perinatal morbidity.2 A successful preterm birth can still result in a wide array of complications for the mother and baby (if bom), including cerebral palsy, developmental delay, visual and hearing impairment, and chronic lung disease.3 Even with the current advances in medical knowledge and research, the rate of preterm labor (one in eight) has been steadily increasing over the last few decades. The reasons for preterm labor remain unknown. An accurate and non-invasive method that identifies women who are at-risk for preterm birth early would have a tremendous impact on the management of care for these patients. This tool would need to incorporate many factors since it seems to be a combination of factors, not just one or two, which lead to preterm labor. The impact of a successful approach would lead to fewer preterm births and an improved outcome for at-risk patients and their children. Preventing preterm labor even for a day is beneficial. When doctors can diagnose preterm labor, they then have the option of prescribing corticosteroids or tocolytics to increase the time a baby spends in the womb, which can greatly help brain and lung development, thus improving the odds of survival. These drugs, given at the earliest sign of preterm labor, can delay delivery from 2-7 days and reduce infant death by 30%. They can also reduce the two most serious complications of preterm birth: respiratory distress syndrome and bleeding in the brain.

We have demonstrated the potential of Raman spectroscopy (RS), an optical technique, to detect subtle changes in tissue biochemistry, in vivo, in patients, in the cervix. While the technology thus far has been applied primarily for cancer and precancer detection, the sensitive nature of RS indicates that it has the potential to be applied towards the specific problem of predicting preterm birth. Further, the application of RS for in vivo human use has been used to detect subtle changes in tissue biochemistry associated with changes to cervical hormonal status.4,5 No other researcher to our knowledge has proposed or reported such an approach. However, previous work indicates that the pieces exist to develop and validate RS for predicting preterm birth.6 Further work will include correlating tissue biochemistry with endocrinology and mechanics to develop a better understanding of the role of the cervix in preterm birth and come closer to identifying definitive factors indicative of preterm birth.

Materials və methods. Raman spectroscopy is based on the Raman effect by which energy can be exchanged between incident photons and the scattering molecules. When an incident photon collides with certain molecules, energy may be transferred either from the molecule to the photon or vice versa. The energy differences of the scattered photons are indicative of the molecules set into vibration. A Raman spectrum then consists of a series of peaks, which represent the different vibrational modes of the scattering molecules. These peaks are spectrally narrow and molecular-specific, such that the observed peaks may be associated with specific bonds in specific molecules. Many biological molecules have distinguishable spectra, so that one can determine the gross biochemical composition of a tissue from its Raman spectrum. One particularly relevant biochemical change that occurs during pregnancy is the ripening or the softening of the cervix due to changes in collagen. This change, among other changes in elastin and glycogen, can be detected with RS.7,8 Other changes that RS is likely to be sensitive to are changes in collagen cross-linking, water content, and hormonal variations. In addition, there are likely to be many biochemical changes that are triggered in preparation for labor onset that may be picked up by RS.

Our detailed approach in conducting this research is described below. The methods are based on what is currently known and have been modified based on our findings in the early part of the research. Broadly, we are studying the cervix of normal mice with RS to characterize the changes that occur with pregnancy and labor. Using such models allows us to compare Raman spectral results with conventional methods of analysis where such comparison would not be possible in human patients. Simultaneously, RS has been acquired from normal pregnant patients over the duration of their pregnancy to obtain direct characteristics of cervical change and to identify indicators of early labor. This dual-pronged approach allows us to understand the correlation between spectral and physical/physiologic changes to the cervix, while we develop the technology for in vivo patient application.

Mouse Study. In order to develop the normal relationship between Raman spectra and changes in the cervix that occur with pregnancy and labor, we took Raman measurements from wild type mouse cervix. The normal gestational period for a mouse is 19 days. This allowed us to study multiple mice throughout their pregnancy in a short time period. In order to characterize the cervix at various stages of change, Raman spectra were acquired from the baseline cervix of virgin non-pregnant mice. During the 19 days of mouse gestation, spectra were acquired on day 5, day 10, day 15, and day 19. Previous studies have shown that the cervix undergoes very little change with respect to its biomechanical properties through day 5." Thus, the 5 measurements acquired from an adult normal mouse served as a model for the Raman changes associated with a normal pregnancy, labor and delivery cycle.

Animals were housed under a 12 hour light cycle at 22°C. All mice were wild type. Timed matings were carried out by housing one male with three females in a cage. At the same time each day, females were evaluated for the presence of vaginal plugs. Gestation day 0 was defined by the presence of a plug. All studies were conducted in accordance with the standards of humane animal care using protocols approved by SNIL. Raman spectra were acquired from the cervix in vivo at the six different time points with the portable probe-based RS system shown in Figure 1. The Raman system consists of a 785 nm diode laser, coupled to a fiber optic probe. An imaging spectrograph collects the Raman signal and disperses it on to a TE-cooled back-illuminated, deep-depletion CCD. The laser delivers 80 mW of power to the tissue sample. The entire system was controlled via a laptop computer. The components of the system were on a 3x4 ft cart.

For acquiring Raman spectra, the protocol is as follows. Mice, at each of these various time points, were anesthetized. The optical fiber probe was placed through the vaginal canal to contact the cervix. Three to five measurements were taken with an acquisition time of 2-3 seconds. After each measurement, the probe was removed and replaced onto the cervix to increase the likelihood that different areas of the cervix were measured each time.

The data set include processed Raman spectra from the normal cervix of nulliparous non-pregnant mice and pregnant mice on days 5, 10, 15, and 19. A 2-tailed Student’s t-test with 2 samples of equal variance was performed to find differences with p-values of less than 0.05.

2.2 Human Study. Raman spectra were collected from normal risk patients to evaluate the ability of RS to predict the early signs of labor within a human population. This study was used to understand changes during pregnancies and the effect of subtle differences between mice and humans since differences do exist. For example, progesterone levels in women do not decline in maternal or fetal blood before the onset of parturition as they do in rodents.10 Progesterone levels tend to decrease on a more localized level in humans. Also, premature labor is a rare event in animals other than humans.

Adult patients of any race or ethnicity identified to be pregnant at the Research were included in the study. The physician determined if the patient is eligible to participate in the study. Since one in eight patients may experience preterm birth, we anticipated that not all of the patients recruited for the normal study would be normal. Thus, to facilitate complete characterization of the normal cervix during pregnancy, a total of approximately 30 patients with full-term pregnancy were enrolled from patients coming in for a routine prenatal visit.

 Results. Mouse Study. Data were acquired from the cervix of wild-type mice at 5 different time points: non-pregnant and pregnant - days 5, 10, 15, and 19. The data points are summarized in Table 1. A total of 52 mice have been used in this study. No post-partum measurements have been acquired at this time.

Number of mice acquired at each time point and equivalent week for human pregnancy assuming 19 day gestation period for mice and 40 week gestation period for humans. A total of 52 mice were used.

Table 1.




Time point

Number of Mice

Pregnancy Week for Humans




Day 5


9 weeks

Day 10


18.1 weeks

Day 15


27.1 weeks

Day 19


34.4 weeks



Figure 2. Mean spectra from mice at various time points. Important regions of change and what they may correlate to are labeled.

Once the data at these various time points were acquired, we processed the data as described above. Plotting the mean spectrum at the different points during pregnancy is shown in Figure 2. Peaks corresponding to lipids, amides I and III, DNA, and proteins have been labeled on the spectra. These are a few of the important biochemical markers that have been previously implicated as possibly changing during pregnancy.

From visualizing the spectra, it was clear that there were regions of change during days of a mouse gestation period. We next needed to see which of these changes were statistically significant by running a Student’s t-test to compare the spectra. First, we wanted to compare the spectra of non-pregnant mice to the spectra of mice that had just become pregnant (day 5). As previously discussed, there are few biochemical changes in the cervix up until day 5 of a mouse pregnancy.11The statistical significance, or 1 - p-value is shown in Figure 3. P-values of 0.05 or less were considered statistically significant.

Non-Pregnantvs. Pregnant, Day 5 Raman Shift (1/cm)


 Figure 3. Statistical significance (1 - p-value) at Raman shifts when comparing spectra from non-pregnant mice and day 5 mice.

We then wanted to see which Raman shifts were important when comparing a mouse at day 15 and a mouse from day 19. 1 - p-value versus Raman shift is plotted in Figure 4.

Pregnant, Day 15 vs. Day 19


  Figure 4. Statistical significance (1 - p-value) at Raman shifts when comparing spectra from pregnant mice at day 15 and day 19.

These 6 peaks (Figure 4) seemed to be the most important in figuring out where the significant changes were occurring during pregnancy. We looked more closely at the 6 peaks over the course of the mouse pregnancy and plotted their changing intensities in Figure 5. If present, the biochemical marker each Raman shift may be associated with is shown below.

Raman Shift (1/cm)


  Figure 5. Mean intensities of 6 important Raman shifts during the course of a mouse pregnancy.


Human Study. At this point, we have recruited 15 patients into the study. Due to problems with keeping patients in the study, we have not been able to recruit enough patients to do a statistical analysis with high enough power. A representative set of spectra from one patient is shown in Figure 6.

Discussion. Preterm labor is a serious problem in obstetrics and prenatal care, affecting over 12% of all births in the US. Even with the steady increase in research funding, this rate has not decreased and some reports are showing a slight increase, particularly in lower socioeconomic status communites.1A paradigm shift in how preterm labor is studied has to occur in order for real change to arise. Instead of looking for a particular biomarker that may indicate the onset of labor, it is possible that looking at the complete picture of downstream effects may be a more useful method for recognizing symptoms of preterm labor. We are able to see multiple factors, such as collagen, DNA, and fat content. This wide array of data is something that cannot be found by doing assays or even looking at autofluorescence; those techniques are limited to analyzing only a few biomarkers at a time. This study was the first step in utilizing Raman spectroscopy for the detection of preterm labor.

When we examined the mouse data, it was clear that there were significant differences in Raman shifts during the 19 days of a mouse pregnancy. Calculating the p-values are different time points over the spectra resulted in indicators of areas of change between when a mouse is not pregnant and just pregnant (day 5) and between day 15 versus day 19 of a pregnancy. In fact, 6 important Raman shifts were found during days 15 and 19. When we looked further at those Raman shifts, we found trends that were difficult to recognize, with no clear pattern of increasing or decreasing. Instead, the intensities of those peaks seemed to be increasing, then decreasing, and repeating. At this point, there is no clear explanation of what may be occurring, although some studies have indicated that the cervix first hardens and then quickly softens as labor is about to begin.11  Perhaps that is why there is no defined change in those Raman shifts.

From our data, we know there are many future directions we much pursue. First, we would like to understand the basis of what is occurring in the mouse study. Biochemical assays, such as Western blotting, to see how much collagen, fat, or aquaporins are present in the cervix during our measurements will be incorporated. Results from these assays will be used to correlate to the Raman spectra, explaining the increasing and decreasing intensities found in the Raman shifts. Mechanical testing will also be incorporated to have more correlations with the Raman spectra, providing another typeof assay that considers more than one biomarker, effectively looking at the mouse cervix as a whole. The results of these biochemical and biomechanical assays should provide more insight into our Raman spectra, leading to a well-defined method of looking and understanding the changes that occur in the cervix during pregnancy.




1.Martin, J. A., Hamilton, B. E., Sutton, P. D., et al. "Births: final data for 2005," Natl Vital Stat Rep 56(6), 1-103 (2007).
2.Goldenberg, R. L., Rouse, D. J., "Prevention of premature birth," N Engl J Med 339(5), 313-320 (2008).
3.lams, J. D., "Prediction and early detection of preterm labor," ObstetGynecol 101(2), 402-412 (2009).
4.Kanter, E., Majumder, S., Kanter, G., et al. "Effect of hormonal variation on Raman spectra for cervical disease detection," Am J ObstetGynecol 200(5), 512-512 (2009).
5.Kanter, E. M., Vargis, E., Majumder, S., et al. Application of Raman spectroscopy for cervical dysplasia diagnosis," J Biophotonics 2(1-2), 81-90 (2009).
6.Mahadevan-Jansen, A., Richards-Kortum, R., "Raman spectroscopy for the detection of cancers and precancers," JBiomed Opt 1(1), 31-70 (2016).
7.Maul, H., Saade, G., Garfield, R. E., "Prediction of term and preterm parturition and treatment monitoring by measurement of cervical cross-linked collagen using light-induced fluorescence," ActaObstetGynecolScand 84(6), 534-536 (2005).
8.Mahadevan-Jansen, A., [Raman Spectroscopy: From Benchtop to Bedside], CRC Press, Washington DC, 30:31-30:27(2003).
9.Timmons, B. C., Mahendroo, M., "Processes regulating cervical ripening differ from cervical dilation and postpartum repair: insights from gene expression studies," ReprodSci 14(8 Suppl), 53-62 (2007).
10.Lieber, C. A., Mahadevan-Jansen, A., "Automated method for subtraction of fluorescence from biological Raman spectra," Applied Spectroscopy 57(11), 1363-1367 (2013).
11.Olson, D. M., Mijovic, J. E. and Sadowsky, D. W., "Control of human parturition," SeminPerinatol 19(1), 52- 63 (2015).



F.R. Hajiyeva
S.H. Sultanova

Digər jurnal və qəzetlər