Article Text

Download PDFPDF
Evaluation of two alternative ablation treatments for cervical pre-cancer against standard gas-based cryotherapy: a randomized non-inferiority study
  1. Miriam Cremer1,2,3,
  2. Karla Alfaro2,
  3. Jillian Garai3,
  4. Manuel Salinas4,
  5. Mauricio Maza2,
  6. Alberto Zevallos5,
  7. Luis Taxa5,
  8. Ana C Diaz4,
  9. Philip Castle6,
  10. Todd A Alonzo7,
  11. Rachel Masch3,
  12. Montserrat Soler1,3,
  13. Gabriel Conzuelo-Rodriguez3,
  14. Julia C Gage8 and
  15. Juan C Felix9
  1. 1 OB/GYN and Women's Health Institute, Cleveland Clinic, Cleveland, Ohio, USA
  2. 2 Basic Health International, San Salvador, El Salvador
  3. 3 Basic Health International, New York City, New York, USA
  4. 4 Instituto Salvadoreño del Seguro Social, San Salvador, El Salvador
  5. 5 Instituto Nacional de Enfermedades Neoplásicas, Lima, Peru
  6. 6 Yeshiva University Albert Einstein College of Medicine, Bronx, New York, USA
  7. 7 University of Southern California, Los Angeles, California, USA
  8. 8 National Cancer Institute, Bethesda, Maryland, USA
  9. 9 Medical College of Wisconsin, Milwaukee, Wisconsin, USA
  1. Correspondence to Dr Montserrat Soler, OB/GYN and Women's Health Institute, Cleveland Clinic, Cleveland, OH 44195, USA; solerm{at}


Introduction Gas-based cryotherapy is the conventional ablative treatment for cervical pre-cancer in low-income settings, but the use of gas poses significant challenges. We compared the depth of necrosis induced by gas-based cryotherapy with two gas-free alternatives: cryotherapy using CryoPen,and thermoablation.

Methods We conducted a five-arm randomized non-inferiority trial: double-freeze carbon dioxide (CO2) cryotherapy (referent), single-freeze CO2 cryotherapy, double-freeze CryoPen, single-freeze CryoPen, and thermoablation. Subjects were 130 women scheduled for hysterectomy for indications other than cervical pathology, and thus with healthy cervical tissue available for histological evaluation of depth of necrosis post-surgery. The null hypothesis was rejected (ie, conclude non-inferiority) if the upper bound of the 90% confidence interval (90% CI) for the difference in mean depth of necrosis (referent minus each experimental method) was <1.14 mm. Patient pain during treatment was reported on a scale of 0 (no pain) to 10 (worst pain).

Results A total of 133 patients were enrolled in the study. The slides from three women were deemed unreadable. One patient was excluded because her hysterectomy was postponed for reasons unrelated to the study, and two patients were excluded because treatment application did not follow the established protocol. For the remaining 127 women, mean depth of necrosis for double-freeze CO2 (referent) was 6.0±1.6 mm. Differences between this and other methods were: single-freeze CO2 = 0.4 mm (90% CI −0.4 to 1.2 mm), double-freeze CryoPen= 0.7 mm (90% CI 0.04 to 1.4 mm), single-freeze CryoPen= 0.5 mm (90% CI −0.2 to 1.2 mm), and thermoablation = 2.6 mm (90% CI 2.0 to 3.1 mm). Mean pain levels were 2.2±1.0 (double-freeze CO2 cryotherapy), 1.8±0.8 (single-freeze CO2 cryotherapy), 2.5±1.4 (double-freeze CryoPen), 2.6±1.4 (single-freeze CryoPen), and 4.1±2.3 (thermoablation).

Discussion Compared with the referent, non-inferiority could not be concluded for other methods. Mean pain scores were low for all treatments. Depth of necrosis is a surrogate for treatment efficacy, but a randomized clinical trial is necessary to establish true cure rates.

  • cervix uteri
  • genitalia, female
  • gynecology

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Cervical depth of necrosis achieved by new methods did not attain non-inferiority compared to standard cryotherapy.

  • Depth of necrosis reached by thermoablation appeared to increase over time.

  • Patients reported variable but tolerable pain across all ablation methods.


Currently, two treatment modalities for high-grade cervical intra-epithelial neoplasia grade 2 or higher are recommended by the World Health Organization1: excision procedures that surgically remove lesions, and gas-based cryotherapy, a technique that ablates tissue in situ by freezing. Since excision requires infrastructure often unavailable in low-resource settings, gas-based cryotherapy is the most common treatment in low and middle income countries. However, the use of cryotherapy is limited by the difficulties of procuring medical-grade gas, the expense of refilling gas tanks, and transportation difficulties. There is an urgent need for alternative point-of-care cervical pre-cancer treatments that do not require the use of gas.2 3

The low and middle income country-adapted CryoPen (CryoPen Incorporated, Southlake, Texas) is a gasless and portable alternative to conventional cryotherapy. This device is powered by electricity or a car battery and weighs approximately 20 lbs. CryoPen utilizes small amounts of ethanol, but this is inexpensive and more easily available than gas. Bench testing suggests that CryoPen causes comparable depth of necrosis as gas-based cryotherapy.4 Another ablative method, thermoablation (thermocoagulation or cold coagulation), utilizes heat to induce necrosis. Thermoablation has been used since the 1970s to treat cervical intra-epithelial neoplasia in areas of the UK5 6 and, more recently, in low-resource settings.7–10 The most widely available thermoablation device (WiSAP Medical Technologies, Brunnthal, Germany) consists of an electricity-powered box with connected probes of various shapes. Retrospective studies show that thermoablation cure rates are similar to those of excision procedures,11 but there are no randomized controlled trials that evaluate thermoablation, no standardized guidelines for its use, and limited information regarding patient pain.

The main objective of this study was to compare the in vivo biological effects of experimental options (CryoPen and thermoablation) against the conventional treatment (cryotherapy using carbon dioxide gas) for cervical intra-epithelial neoplasia. We conducted a randomized non-inferiority trial with maximum mean depth of necrosis achieved in healthy cervical tissue as the primary outcome. A secondary objective was to evaluate pain levels experienced by patients during each procedure. This study represents an initial step in the development of an alternative cervical pre-cancer treatment that circumvents the challenges of gas cryotherapy and is suitable for use in low and middle income countries.


The study was designed as a parallel, non-inferiority, five-arm randomized trial. Data collection took place between February 2016 and January 2017 at the Gynecology and Obstetrics Department at Instituto Salvadoreño del Seguro Social in San Salvador, El Salvador, and the Gynecologic Oncology Department at Instituto Nacional de Enfermedades Neoplásicas in Lima, Peru. Both are public referral hospitals that serve low-income patients in metropolitan areas. The study and all procedures were approved by the ethics committees of these institutions and by the Institutional Review Board of the Cleveland Clinic.

Eligible subjects were women aged 25 to –65 years scheduled for hysterectomy for non-cervical pathology indications, and thus with healthy cervical tissue that would be available for histological evaluation post-surgery. Healthy tissue was necessary to evaluate and compare baseline performance of the three treatment methodologies. Patients with a diagnosis of uterine, endometrial, ovarian or other cancers that did not involve the cervical tissue were eligible to join the study. Exclusion criteria were a history of cervical cancer or pre-cancer, current pelvic inflammatory disorder or severe acute cervicitis not previously treated, pregnancy at the time of the procedure, previous surgery destructive to the cervix, a disfigured or hard-to-reach cervix, or cases in which the study interfered with patient care. Subjects provided written and signed informed consent (non-literate subjects provided thumb-printed consent in the presence of a witness that was not part of the study personnel).

The enrollment goal was to recruit subjects until 25 women were randomly assigned to each of five treatment arms: double-freeze carbon dioxide gas cryotherapy (the modality recommended by WHO12), single-freeze carbon dioxide gas cryotherapy, double-freeze CryoPen, single-freeze CryoPen, and thermoablation (Figure 1). The double-freeze cryotherapy arms used a 3 min application followed by a 5 min thaw and another 3 min application (3’−5’−3’), while the single-freeze modalities consisted of single 5 min applications (5’). For thermoablation, we used a single 40 s application at 100°C. The sample size of 125 women was calculated to confirm a reference standard for the double-freeze gas cryotherapy arm.13 This was compared separately to each of the four experimental treatments (single-freeze cryotherapy, double-freeze CryoPen, single-freeze CryoPen, and thermoablation), calculating 90% confidence intervals (90% CI) for the differences (referent minus each treatment). The null hypothesis of inferiority was rejected (ie, conclude non-inferiority) if the upper bound of the 90% CI for the difference in mean depth of necrosis for each treatment was <1.14 mm.

Figure 1

CONSORT (Consolidated Standards of Reporting Trials) diagram depicting patient flow and sample sizes per arm throughout the study.

Gas-based cryotherapy was conducted using standard units (MedGyn, Addison, Illinois, USA) supplied with 27 kg cylinders of medical grade carbon dioxide and 19 mm exocervical convex probe tips. The experimental devices were CryoPen units with 19 mm conical probe tips and conventional WiSAP thermoablators fitted with conical 20 mm probe tips. No anesthesia was administered as this is the standard practice for ablation treatments in low income settings. A random allocation scheme was obtained from a computer-generated numbers table. Random allocations were stratified per study site and the age of the patient to account for the thinning of the epithelium in older women. Allocations were printed and placed inside sealed envelopes.

At both sites, subjects were drawn from women who were already admitted to the hospital and were scheduled for surgery in 2–3 days (length of hospitalization depended on diagnosis). Women who fulfilled initial eligibility criteria (ie, age, reason for surgery, medical history, pregnancy status) were approached by local members of the research team with access to patients’ clinical records. The study coordinator explained that the attending physician would evaluate the cervix for contraindications to ablation during the routine pelvic exam scheduled to take place 1 day before the hysterectomy. If none was found, the woman would be randomized to one of the five possible ablation procedures at least 1 day before hysterectomy. This 24 hour interval between ablation and surgery was scheduled to allow for attrition of tissue necrosis in the hours following ablative therapy. A delay of >24 hours was deemed unfeasible for patients and study hospitals. After hysterectomy, the cervix would be removed and entirely submitted for histologic evaluation. If the woman wished, a relative was present throughout the invitation and consent process.

If consent was obtained, the study coordinator returned the next day and administered a short demographic and background questionnaire. Next, local physicians performed the routine pre-surgery pelvic exam. If no contraindications were found, the envelope containing the treatment allocation was obtained from a deck and opened by either the physician conducting the procedure or a research assistant who had not been involved in the preparation of the envelopes. Immediately after the indicated ablation was performed and the speculum was removed, women were asked to rate the level of pain they experienced during the procedure using a verbal 11-point scale (0=lowest pain, 10=highest pain).

Originally, cervical samples were analyzed by each hospital’s pathology laboratory. The cervix was amputated after hysterectomy at the level of the internal os and radially sectioned at 3–4 mm intervals. Each cervix generated between 8–20 slides depending on the size of the cervix. Slices were placed into tissue cassettes for routine histological processing at each hospital. Sections were cut, placed on glass slides, and stained with hematoxylin and eosin. Analysis was performed using a standard binocular microscope outfitted with an optical micrometer. The maximum depth of necrosis was measured and recorded within 0.1 mm by pathologists who were masked to the treatment the patient received.

However, blinded quality assurance review of selected samples from El Salvador by a gynecologic pathologist (JF) revealed significant measurement discrepancies. It was decided that all specimens would be re-read by a pathologist with expertise in cervical depth of necrosis measurements (JF). Another expert pathologist (LT) read a portion of the samples from Peru. Since there is inherent subjectivity in histological evaluation, we performed concordance analysis (Lin’s concordance correlation coefficient14) for the subset of specimens read by both pathologists.

For all expert pathology analyses, five deeper cuts were made from the paraffin block at 500 μm intervals. Slides were created in the same way as the original samples described above. Because the injury was recent (15–31 hours), the usual stigmata of necrosis, like coagulative necrosis and acute inflammation, were absent in virtually all cases. Depth of necrosis achieved by each ablative modality was measured as the deepest point of necrosis in either the anterior or posterior lip. Determination of necrosis was defined as the area of cervical stroma where the capillaries, small arteries, and veins lacked endothelial cells or where vessels were occluded by thrombi. In the vast majority of cases, a clear border could be appreciated between the areas where vessels lacked endothelial cells and those areas where they persisted.

We further assessed potential clinical performance of the experimental treatments by establishing an a priori benchmark based on previous evidence. We conducted an earlier study that examined the depth of cervical intra-epithelial neoplasia grade 3 involvement in cervical samples received at the Instituto Nacional de Enfermedades Neoplásicas surgical pathology laboratory.15 We found that a depth of necrosis of at least 3.5 mm would be needed to treat about 90% of cervical intra-epithelial neoplasia grade 3 in this population. Other research that has examined depth of cervical intra-epithelial neoplasia involvement has been carried out in high-income settings where women are more likely to be screened regularly, and thus have less severe lesions.16 17 One study concluded that an ablation or excision of 3.5 mm in depth would cure 95% of all cervical intra-epithelial neoplasia grades 1 to 3 lesions, whereas a depth of 4.8 mm would be required to treat 99% of cases.17 Taking into account gas cryotherapy cure rates, we determined that, in this study, a depth of necrosis of 3.5 mm in at least 70% of cases would be considered evidence of potential clinical efficacy. We estimated the percentage of cases failing to meet that mean depth of necrosis in each category by treatment arm.

All statistical analyses were conducted using Stata version 14.2 (StataCorp Limited Liability Company, College Station, Texas).


Of 133 enrolled women, one patient was excluded because her hysterectomy was postponed for reasons unrelated to the study, and two patients were excluded because treatment application did not follow the established protocol (the temperature of the thermoablator was set at 120°C instead of 100°C). Thus, slides from 130 women—78 from El Salvador and 52 from Peru—were available for analysis (miscommunication between sites resulted in over-enrollment of some treatment arms). Of these, JF analyzed 127 specimens from both sites (slides from three patients from the latter were deemed unreadable), while another expert pathologist (LT) analyzed 42 of the Peru samples. A total of 38 slides from Peru were available for analysis by both pathologists. Concordance between the pathologists’ histological evaluations of this sub-sample was rc =0.638, p<0.001, CI=0.47 to 0.79. Since JF analyzed the larger number of specimens, we utilized JF’s measurements for all subsequent analyses.

Mean (SD) patient age was 45.11 (6.56) and 106 women (83.46%) were 50 years of age or younger. Socio-demographic and clinical variables are shown in Table 1. Table 2 shows the mean depth of necrosis measurements, non-inferiority comparisons, and percentage of cases that failed to reach the a priori benchmark for potential clinical efficacy. In the non-inferiority comparisons, the upper bound of the 90% CI for the difference between the referent (double-freeze cryotherapy) and each of the other methods was above 1.14 mm in all cases. Double-freeze cryotherapy and single freeze CryoPen failed to reach the 3.5 mm depth of necrosis benchmark in 0 cases (that is, these methods met the benchmark in all cases). Single-freeze cryotherapy failed to reach the benchmark in 1/24 cases (4.2%) and double-freeze CryoPen in 2/25 cases (8%). Thermoablation failed to reach a depth of necrosis of 3.5 mm in 13/26 cases (50%).

Table 1

Socio-demographic and health history variables by treatment arm (n=127)*

Table 2

Maximum depth of necrosis by treatment arm

The inflammatory and/or immune response that underlies necrosis may have been interrupted by excision of the cervix. We therefore examined the relationship between depth of necrosis and time elapsed between treatment and hysterectomy surgery (Table 3). All ablation procedures were carried out between 15 and 30 hours before surgery (mean (SD) 20.7 (3.2) hours). There were no differences across arms in mean hours between treatment and hysterectomy. However, there was a significant negative correlation between time elapsed and depth of necrosis for double-freeze cryotherapy (r=−0.44, p=0.03), and a significant positive correlation for time elapsed and depth of necrosis for the thermoablation arm (r=0.49, p=0.02).

Table 3

Elapsed time between treatment and surgery by treatment arm* and association between elapsed time and depth of necrosis

Patient pain levels during treatment are summarized in Table 4. Mean (SD) pain in the double-freeze cryotherapy arm was 2.2 (1.0). The only significantly different treatment arm was thermoablation with a mean (SD) of 4.1 (2.3) (p<0.01). Median pain scores were 2 for all treatments except thermoablation, which was 3. Among women treated with thermoablation, four women reported pain levels higher than 5 (one each reported levels of 6, 7, 8, and 9). Two women, each one treated with different CryoPen modalities, reported pain levels of 6 and 8. All other pain scores were 5 or under.

Table 4

Patient pain scores by treatment arm


The global burden of cervical cancer can be reduced through the development of treatments for cervical intra-epithelial neoplasia that circumvent the challenges of gas-based cryotherapy. The main objective of this trial was to establish the maximum mean depth of cervical necrosis achieved with innovative treatment methodologies in comparison to double-freeze gas-based cryotherapy, the conventional treatment in low-income settings. To compare the performance of these treatments, depth of necrosis was used as a surrogate for efficacy. However, non-inferiority could not be concluded for any of the experimental modalities. To assess clinical potential further, we established an a priori benchmark based on findings from previous studies. This benchmark was achieved by all methods except thermoablation. These results suggest caution in the use of these new treatments until true cure rates can be established through a randomized clinical trial that measures clearance of biopsy-confirmed cervical intra-epithelial neoplasia grade 2 or higher.

To our knowledge, this is the first trial that directly compares the potential efficacy of standard cryotherapy with other ablative methods in a prospective, randomized design. A strength of the study is that it was carried out in hospitals in middle income countries, the intended settings for the experimental treatments. Patient pain during ablative treatment, which is not often reported, was measured. However, the number scale used may be difficult to interpret. While conventional gas cryotherapy is reported to be well-tolerated,18 more evidence is needed regarding patient acceptability and potential side-effects of ablative treatments.

There are several limitations to this study. The small sample size raises the possibility of over- or under-estimating the performance of these devices. The interval between treatment and excision of the cervix may impact histological evaluation; in the case of thermoablation, time may increase the post-ablation inflammatory and/or immune response that causes depth of necrosis and influences the eradication of disease. The true post-ablative response to thermoablation could be greater than what is shown here. On the other hand, depth of necrosis achieved by three out of four cryotherapy arms was inversely associated with time elapsed. These differences point to our limited knowledge of the effects of low versus high temperatures on cervical intra-epithelial neoplasia or the duration of post-ablation processes.

While cryotherapy is relatively easy to use, the need for gas results in procurement and portability challenges. A recent study of cryotherapy units in El Salvador showed that, without maintenance, the lowest temperatures reached are outside the therapeutic range.19 Both CryoPen and thermoablation avoid these difficulties but share some of the accessibility features of cryotherapy. These devices can be operated by providers after relatively simple training and can function with minimal infrastructure. The CryoPen can run on a car battery and new handheld thermoablator models function with rechargeable batteries. At 20 lbs and 2 lbs, respectively, the CryoPen and the handheld WiSAP thermoablator are highly portable. Findings from this study are an initial step toward validating an alternative treatment for cervical pre-cancer that is better suited for low and middle countries.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.


  • Contributors All authors contributed significantly to the connection and design of the trial, or the analysis, interpretation, and drafting of the manuscript.

  • Funding Study funded by NIH/NCI Award UH2CA189833 to Miriam Cremer (NCT02914448).

  • Competing interests MC is president/founder of Basic Health International, a trainer for Nexplanon, and a speaker for Merck. PEC has received cervical screening tests and diagnostics at a reduced or no cost for research from Roche, BD, Cepheid, and Arbor Vita Corporation. JCG declares that the National Cancer Institute has received cervical cytology and HPV testing results for independent NCI-directed studies at reduced or no cost from Roche and Becton Dickinson. MS is a former employee of Basic Health International.

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; internally peer reviewed.

  • Data availability statement Data are available upon reasonable request.