Dr C Founta, Department of Obstetrics and Gynaecology, University Hospital of Ioannina, Stavrou Niarchou 1, 45500, Ioannina, Greece. Email firstname.lastname@example.org
Please cite this paper as: Founta C, Arbyn M, Valasoulis G, Kyrgiou M, Tsili A, Martin-Hirsch P, Dalkalitsis N, Karakitsos P, Kassanos D, Prendiville W, Loufopoulos A, Paraskevaidis E. Proportion of excision and cervical healing after large loop excision of the transformation zone for cervical intraepithelial neoplasia. BJOG 2010;117:1468–1474.
Objective To determine how the proportion of the cervical volume excised affects cervical regeneration.
Design Prospective observational study.
Setting University Hospital.
Population Women planning to undergo excisional treatment for cervical intraepithelial neoplasia who wish to have future pregnancies.
Methods The cervical volume (and dimensions) is calculated with magnetic resonance imaging (MRI) before treatment. The volume (and dimensions) of the cone is assessed before fixation by a volumetric tube and a ruler; the percentage (%) of excision is computed. Cervical regeneration is estimated by repeat MRI at 6 months.
Main outcome measures Cervical regeneration in relation to proportion of excision. Statistical analysis was performed by box plots and analysis of variance.
Results A total of 48 women have been recruited; 29 have completed 6 months follow up. Both the total cervical volume (from MRI) before treatment and the volume of the excised/ablated cone varied substantially. The estimated proportion of excision varied significantly between 4% and 39% (median 11%). Multivariate linear regression revealed that the proportional deficit at 6 months post-treatment was determined mainly by the proportion of the excised volume.
Conclusions Careful assessment of risks and benefits of treatment is essential when deciding to treat women who wish to have future pregnancies. Assessment of the proportion of the cervical volume and length excised might identify those that need further surveillance during future pregnancy.
Treatment for cervical intraepithelial neoplasia (CIN) should not only be effective in eradicating lesions but should also have minimum morbidity. Meta-analyses have shown no obviously superior surgical technique among excisional and destructive methods in terms of failure rates,1 risk of future invasive disease2,3 and both short- and long-term complications.4 Large loop excision of the transformation zone (LLETZ) is by far the most popular method because it allows a comprehensive evaluation of the excised tissue, while being relatively easy, simple, quick, cheap and acceptable by women.
Although intraoperative and postoperative short- and long-term complications were thought to be relatively mild and rare, recent meta-analyses5,6 and large linkage studies7–14 have revealed that excisional techniques, in particular, are associated with future perinatal morbidity. Cold knife and laser conisation as well as LLETZ increased the risk of prematurity and low weight at birth;5 in addition an association between LLETZ and the risk of preterm prelabour rupture of membranes was shown.5 Serious clinical outcomes, such as severe (<32–34 weeks of gestation) or extreme (<28–30 weeks of gestation) prematurity and perinatal mortality were also more common in women with a history of laser or cold knife conisation.6 Laser ablation had no adverse effect.
Inevitably the amount of tissue excised or destroyed is an important risk factor. This might be more significant than the actual method used. LLETZ can remove a variety of different cone sizes and if only large cones were to be included, even this method would probably also be related to serious adverse effects. The analysis of three studies that provided data on the cone depth revealed increased risk of prematurity only when the depth exceeded 10 mm.5 Recent data from Denmark support an increase in risk by increasing cone depth and a critical cutoff that signifies a rapid rise in prematurity rates.15
Conversely, if the volume of the cervix and the height of the endocervical canal vary significantly across individuals then similar-sized cones would represent various proportions of excision in different women. Adverse outcomes are most probably related to the proportion of the cervical volume and endocervical canal excised rather than the actual depth of the excision or the method used.
Although strong, the current available evidence is retrospective; most data derive from cohort or linkage studies and are therefore prone to selection or reporting bias and insufficiently controlled for confounding factors. A prospective study assessing the proportion of cervical volume or endocervical canal excised and its effect on cervical healing as well as future pregnancy has never been conducted.
This prospective observational study aims to evaluate which factors best predict the regeneration of the cervix: the absolute dimensions of the cone or the proportional volume of excised or ablated tissue.
Inclusion, exclusion criteria and outcomes
This was a prospective observational study conducted in the University Hospital of Ioannina in Greece between January 2009 and February 2010. All women aged 23–42 years who planned to undergo excisional treatment for CIN and wished for future pregnancies were included. Women were included irrespective of the grade of the lesion and parity. Women with previous cervical treatment and those that had completed their family were excluded.
We assessed the variation in cervix size as well as the cervical volume before and after treatment and determined the relation between the deficit of the regenerated cervix with the dimensions of the cone, expressed both in absolute terms and in relative terms (compared with the original cervix).
The dimensions of the cervix before treatment were assessed using transverse T1-weighted, transverse and sagittal T2-weighted magnetic resonance imaging (MRI). The volume was computed using the cylinder formula16,17 Volume = π [(anteroposterior diameter + transverse)/4]2 × length]. The distance between the external and internal os was defined as length. The centre of this line was considered the midpoint of the cervical canal and was used to draw the anteroposterior and transverse diameters of the cervix (Figures 1 and 2).
Excisional treatment was performed by an experienced colposcopist. The volume of the cone was assessed before fixation by submergence in a fluid-filled cylindrical vial. The displaced volume of fluid was computed as (difference in height × circular surface of the base of the vial). The height and ectocervical diameter of the cone were also recorded. The ball diathermy used to achieve haemostasis can occasionally add significant tissue damage. This additional tissue loss was subjectively assessed by the colposcopist (expressed as the proportion of the cone volume excised) and added to the volume of the cone. The size of the lesion in terms of involved quadrants and the type of the transformation zone were also recorded.
The MRI was repeated during the first follow-up visit at 6 months to assess cervical healing. The post-treatment cervical volume was calculated and correlated to the proportion of volume excised.
We computed the mean, median and quartiles, plotted histograms and assessed the variations across individuals for various variables, such as the volume of the cone, the volume of the cervix pre-treatment and at 6 months follow-up, the proportion of the volume excised (volume excised cone/volume cervix before excision), the deficit (volume before excision—volume 6 months after excision) and the proportional deficit at 6 months (volume at 6 months/original volume of the cervix). The relation between the cone volume and the grade or the size of the lesion (number of quadrants involved) was assessed by box plots, a trend test18 and analysis of variance. We ultimately attempted to predict the proportional volume deficit at 6 months using multivariate linear regression with the following covariates: age, parity, grade and size of the lesion, original absolute cervical volume, cone volume and proportion excised.
A total of 48 women, all treated by LLETZ, have been recruited. Twenty-nine have completed 6 months of follow up. Age varied between 23 and 42 years (median = 31.5 years); 32 (67%) were nulliparous. The majority had a high-grade lesion on histology: 5 (10%) CIN1, 25 (52%) CIN2 and 18 (38%) CIN3. The number of quadrants involved was one in four (8%) women, two in 12 (25%) women, three in eight (17%) women and four in 24 (50%) women, respectively.
There was a positive and statistically significant correlation between the number of involved quadrants and the severity of CIN (chi square test for trend; see Table 1).
Table 1. Relation between size (involved quadrants) and severity of CIN
Number of quadrants affected by CIN
Severity of CIN
Row proportions are given in italics.
The cervical volume (by MRI) before treatment varied substantially between individuals from 11 to 40 cm3 (median 23 cm3) and the volume of the excised/ablated tissue ranged between 1 and 7 cm3 (median 3 cm3). The estimated percentage of excision varied between 4% and 39% (median 11%). The distributions of the cone dimensions (height and volume), the height and volume of the cervix before and 6 months after treatment as well as the range for deficit and proportional deficit at 6 months are shown in Table 2 and in Figure 3.
Table 2. Distribution of the dimensions of the excised cone (± volume ablated), and those of the cervix before and 6 months after LLETZ
SD, standard deviation of the mean; Q1, first quartile; Q3, third quartile.
Anteroposterior width (mm)
Transversal width (mm)
Volume cone excised (cm3)
Volume cone + ablated tissue (cm3)
Proportion excised (%)
Proportion excised + ablated (%)
Volume before LLETZ (cm3)
Volume at 6 months (cm3)
Deficit at 6 months (cm3)
Proportional deficit (%)
The volume of the cone (in absolute volume [cm3] or as a proportion of cervix excised [%]) increased with the number of involved quadrants and increasing severity of CIN (box plots in Figure 4). However, when this was incorporated in a multivariate analysis of variance, it was only the number of involved quadrants that determined the cone volume.
The absolute deficit at 6 months post-treatment correlated positively with the cone volume and negatively with the original volume before LLETZ. However, in a multivariate linear regression, proportional deficit was determined only by the proportion of the excised volume. Age, grade of CIN, number of involved quadrants, absolute pre-treatment volume of the cervix and cone volume did not improve the fit of the linear regression. Increasing the proportional volume excised/ablated by 1% raised the proportional deficit by 0.91% (95% CI 0.65–1.17%).
Four women from the study population became pregnant after treatment. One woman required a second treatment for recurrent/residual high-grade lesion.
Recent evidence from retrospective studies indicates that the height of the cone is a predictor for subsequent prematurity.5,15 However, pre-treatment cervical volumes vary significantly across individuals; consequently, similar cone volumes might represent different proportional deficits and lead to different regeneration patterns. This is the first study reporting the association between the deficit of the regenerated cervix as a relation of the proportional volume excised/ablated using MRI.
There appears to be no explanation for why different treatment methods would result in diverse adverse outcomes. It is most likely that differences across techniques are explained by the proportion of tissue removed or destroyed. Cold knife conisation removes large cones, whereas the colposcopically directed laser beam allows clinicians to destroy less healthy cervical tissue. It might be logical to expect similar serious adverse outcomes if the same proportion of tissue was to be ablated by laser.
The potential relation of the cervical proportional deficit to the risk of prematurity remains to be seen after the completion of the follow-up period. Women who become pregnant have serial cervical length assessments by transvaginal ultrasound and their obstetric outcomes are recorded.8,9,14 Those outcomes will be assessed in relation to the proportion of excision and regeneration as assessed by MRI at 6 months post-treatment.
Mechanisms that lead to prematurity might involve the anatomical or histological alterations of the healed cervix and changes in the cervicovaginal flora. Occurrence of treatment failure (5–10%)19–21 requiring repeat conisation provides a specimen for further assessment of the regenerated tissue (composition and organisation of collagen fibres)22,23 and of the immunological environment (i.e. altered cytokines). It can be hypothesised that beyond a certain cutoff of proportional deficit, not only the regenerated tissue quality is affected but also cervical function and this probably compromises pregnancy outcomes. Such a determination could lead to more targeted interventions in future pregnancies.
Albeit optimally accurate as a research tool, the significant financial cost of MRI restrains its use in clinical practice. Currently, we are comparing the precision of MRI with three-dimensional transvaginal ultrasound measurements of the cervical volume in women immediately before hysterectomy using postoperative anatomical measurement as gold standard. If three-dimensional transvaginal ultrasound appears to yield measurements of acceptable precision, it could be proposed as a feasible assessment procedure in future large-scale studies that could establish the risk of adverse pregnancy outcomes associated with the proportional volume of removed cervical tissue.24
The Hippocrates principle of ‘do no harm’ should rule every medical intervention. The ideal treatment modality should have maximal cure rates with minimal sequelae. Additional interventions that could minimise morbidity should be continuously explored.
Caution is therefore recommended when deciding to treat young women with potentially clinically insignificant mild cytological abnormalities. Until new methods and biomarkers for the triage of lesions with a true cancerous potential become available, appropriate selection is essential. If the decision is made to proceed to treatment, women should be appropriately counselled and every effort should be made to eradicate the lesions without removing excess healthy cervical tissue.
Adverse outcomes are more likely to be related to the proportional deficit rather than to the actual depth of the excision or method used, for this reason pre-treatment and 6-months post-treatment cervical volume assessment using transvaginal ultrasound might successfully predict women at risk.
Disclosure of interests
We declare that we have no conflict of interest.
Contribution to authorship
EP conceived and designed the experiments; CF, GV, MK, AT, PK and EP performed the experiments; MA and CF analysed the data; CF, MA, GV, MK, AT, PLM, ND, PK, DK, WP, AL and EP contributed reagents/materials/analysis tools; and CF, MA, GV, MK, PLM, ND, WP and EP wrote the paper.
Details of ethics approval
This study has been approved by the Ioannina University Hospital’s ethical committee [protocol 28/9-7-2009(Θ.20)].
MA is supported by the Fonds national de la Recherche scientifique ‘TELEVIE’, Brussels, Belgium (ref. 7.4620.09); the Belgian Foundation Against Cancer, Brussels, Belgium; and the European Commission (DG Sanco, Luxembourg) through the ECCG project (European Cooperation on development and implementation of Cancer screening and prevention Guidelines, IARC, Lyon, France).
We thank Dr George Koliopoulos for his comments and the critical appraisal of this manuscript.