The diagnosis of malignant transformation of abdominal endometriosis is still a challenge for gynecologists. There are no characteristic symptoms and markers in the process of carcinogenesis. The best treatment for malignant transformation of abdominal endometriosis is not clear. We searched the articles published from September 1986 to December 2020 by Medline and EMBASE, and the search words were combined into medical subject entries. The keywords used for the search are as follows: "abdominal wall endometriosis" and "clear cell carcinoma". All relevant articles were retrieved and the list of relevant references was systematically reviewed to determine further reports that could be included in this analysis. Besides, reviews of cancer and endometriosis published over the same period were reviewed and their reference lists were searched for potential additional studies. We used SPSS 23.0 statistical software for statistical analysis. The unpaired t-test of continuous variables and the chi-square test of classified variables were used in the comparison between groups. GraphPad's Kaplan-Meier estimation was used for survival analysis.
Finally, we included 45 patients in our systematic retrospective analysis, including 44 from the literature review and 1 from our organization. All of them reported that endometriosis was associated with malignant transformation of the surgical scar. The characteristics of the patients are shown in Table 1. As shown in Fig. 6, it is more common than previously reported, especially if we consider the increasing rate of CS in western countries [9]. The average age of the patients at the time of diagnosis was 46.5 years (range from 37 to 60), which was consistent with the retrospective analysis of Endometriosis-associated malignant transformation in an abdominal surgical scar by Mihailovici et al. [8]. Fifteen cases (33.3%) had a history of endometriosis and 28 cases (62.2%) had never been diagnosed with endometriosis before admission (2 cases were not clear). The average delay from the first operation to diagnosis was 17.9 years (standard deviation 6.6). CCC is related to uterine surgery, mainly the CS. Forty-two cases (91.1%) had at least one CS, of which 15 cases had 2 CS s, 6 cases had 3 CSs, and 4 cases (8.9%) had undergone other gynecological operations, usually CS (Table 2). At least one excision of scar endometriosis was performed in 9 cases, including our case.
Table 1
Summary of studies included in the review.
Author | Year | Age | Delay from first gynecological surgery(yr) | Onset of symptoms(m) | Lump size(cm) | Surgery | Pathology | Patient outcome |
Follow-up time (m) | Relapse | Follow-up time (m) | Death |
Giannella [15] | 2020 | 45 | 15 | 3 | 20 | no | CCC | NA | yes | 7 | yes |
Behbehani[20] | 2019 | 48 | NA | 5 | 7 | yes | CCC | NA | NA | NA | NA |
Tsuruga [12] | 2019 | 49 | 15 | 3 | 5 | yes | CCC + EC | 4.5 | no | NA | NA |
Rivera [19] | 2019 | 48 | NA | NA | 7 | yes | CCC | 2 | no | NA | NA |
Lopes [10] | 2019 | 48 | 12 | 84 | 12 | yes | CCC | 3 | no | NA | NA |
Lai [23] | 2019 | 56 | NA | 396 | 6.5 | yes | CCC | 3 | yes, Inguinal LN | 11 | yes |
Lai [23] | 2019 | 56 | NA | 252 | 12 | yes | CCC | NA | NA | 5 | yes |
Lai [23] | 2019 | 45 | NA | 240 | 4.8 | no | CCC | 7 | yes, Abd. wall, inguinal LN | 7 | yes |
Gentile[26] | 2018 | 42 | 7 | 8 | 10.6 | yes | CCC | 2 | no | NA | NA |
Mihailovici [8] | 2017 | 47 | 22 | 264 | 11 | yes | CCC | NA | NA | NA | NA |
Wei [24] | 2017 | 46 | 18 | 2 | 9.5 | yes | CCC | 3 | no | NA | NA |
Marques [27] | 2017 | 47 | 24 | 3 | 8 | yes | CCC | 45 | no | NA | NA |
Kostrzeba [28] | 2017 | 58 | 38 | NA | 25 | yes | CCC | 3 | no | NA | NA |
Graur[29] | 2017 | 43 | 22 | 7 | 8.5 | yes | CCC | 11 | no | NA | NA |
Ferrandina [16] | 2016 | 44 | 9 | 8 | 22 | yes | CCC | 5 | yes, liver | 6 | yes |
Sosa-Durán [30] | 2015 | 45 | NA | 8 | 9 | yes | CCC | 16 | no | NA | NA |
Ruiz [11] | 2015 | 41 | 20 | NA | 14.8 | yes | CCC | 6 | yes, local | NA | NA |
Ruiz [11] | 2015 | 57 | 30 | 9 | 19.4 | yes | CCC | NA | no | NA | NA |
Aust [31] | 2015 | 47 | 16 | 6 | 10 | yes | CCC | 10 | no | NA | NA |
Liu[32] | 2014 | 39 | 10 | 60 | 6 | yes | CCC | 10 | yes, local | 12 | yes |
Heller [33] | 2014 | 37 | 8 | 96 | 18 | yes | CCC | 5 | yes, | NA | NA |
Dobrosz [34] | 2014 | 42 | NA | 16 | NA | yes | CCC | NA | NA | NA | NA |
Ijichi [35] | 2014 | 60 | 37 | 48 | 4 | yes | CCC | 8 | yes, local | 23 | no |
Shalin [36] | 2012 | 47 | NA | 10 | 3 | yes | CCC | 7 | no | NA | NA |
Sawazaki [17] | 2012 | 41 | 18 | NA | 4.8 | yes | CCC | 4 | no | NA | NA |
Mert [37] | 2012 | 42 | NA | NA | 17.8 | yes | CCC | 26 | no | NA | NA |
Mert [37] | 2012 | 51 | 8 | 12 | 7 | yes | CCC | 49 | no | NA | NA |
Li [38] | 2012 | 49 | 26 | 25 | 9 | yes | CCC | 8 | no | NA | NA |
Yan [21] | 2011 | 41 | 5 | 4 | 6.3 | yes | CCC | 24 | no | NA | NA |
Bourdel [39] | 2010 | 43 | 20 | 60 | 9 | yes | CCC | 6 | yes | 22 | yes |
Williams [40] | 2009 | 53 | 24 | 4 | 4.7 | yes | CCC | 3 | yes, local, InguinalLN, lung | 11 | yes |
Matsuo [41] | 2009 | 37 | 10 | 5 | 14 | yes | CCC | 24 | yes, local | NA | NA |
Rust [42] | 2008 | 42 | NA | 24 | 4.9 | yes | CCC | NA | NA | NA | NA |
Barts [18] | 2008 | 38 | 13 | 151 | 11 | yes | CCC | 8 | yes | NA | NA |
Achach [43] | 2008 | 49 | 20 | NA | 8.5 | yes | CCC | 6 | yes, bladder and pelvic bone | NA | NA |
Razzouk [22] | 2007 | 46 | 26 | NA | 17 | yes | CCC + EC | 3 | yes, liver | 6 | yes |
Harry [44] | 2007 | 55 | NA | 15 | 4 | yes | CCC | 18 | no | NA | NA |
Sergent [45] | 2006 | 45 | 25 | 17 | 20 | yes | CCC | NA | yes | 6 | yes |
Alberto [46] | 2006 | 38 | 11 | 6 | 6 | yes | CCC | NA | NA | NA | NA |
Ishida [47] | 2003 | 56 | 24 | 7 | 10 | yes | CCC | NA | yes, lung, bone and brain | 48 | yes |
Park [48] | 1999 | 56 | 24 | NA | 5 | yes | CCC | NA | NA | NA | NA |
Miller [49] | 1998 | 38 | 9 | 8 | 4 | yes | CCC | 60 | no | NA | NA |
Hitti [50] | 1990 | 46 | 14 | NA | 6 | yes | CCC | 30 | no | NA | NA |
Schnieber Agner-Kolb [51] | 1986 | 40 | 15 | 18 | NA | yes | CCC | NA | yes | 18 | yes |
our case | 2020 | 49 | 19 | 204 | 6.3 | yes | CCC | 1 | no | NA | NA |
Table 2
Characteristics of patients.
| | n | Percentage |
Age(yr) | 46.5 ± 4.9 |
31–40 | 7 | 15.22% |
41–50 | 28 | 60.87% |
51–60 | 10 | 21.74% |
Previous gynecological surgery | 1C | 20 | 44.44% |
2C | 15 | 33.33% |
3C | 6 | 13.33% |
other surgeries | 4 | 8.89% |
Delay from fist gynecological surgery (yr) | 17.9 ± 6.6 |
0–10 | 7 | 20.59% |
11–20 | 15 | 44.12% |
21–30 | 10 | 29.41% |
31–40 | 2 | 5.88% |
Onset of symptoms(yr) | 4.4 ± 5.2 |
0–1 | 19 | 52.78% |
1–2 | 6 | 16.67% |
2–10 | 6 | 16.67% |
10–33 | 5 | 13.89% |
Follow-up time (m) | 15.0 ± 11.5(1–60) |
Because some literature did not specify the time of the events, the results of the delay from fist gynecological surgery,onset of symptoms, and follow-up time only counted the cases with specific years. |
As for tumor markers, CA 125 was detected in 24 cases before the operation. 14 cases were in the normal range (0-35U/mL), 10 cases were higher than the normal range, and the highest was 3157.9U/mL [10]. The preoperative detection of CEA,7 was in the normal range in 9 cases, above the normal range in 2 cases, up to 96.5 µg/L [11]. CA 199 was detected in 12 cases before treatment, and above the normal range in 4 cases, up to 222 µg/L [12]. (Table S1).
Previously, only 15 patients had a history of endometriosis (44 patients had available information) and 12 patients had no abdominal masses before diagnosis without any intermittent, periodic, or persistent pain. But all patients had palpable abdominal surgical scar masses with an average diameter of 10.0 cm (standard deviation 4.4 cm), up to 25 cm. The pathological results of the mass are shown in Table 3. The most common histological type is CCC, in 43 cases (95.6%), followed by CCC with endometrioid carcinoma in 2 cases (4.4%). Of these, only 22 cases (51.2%) found coexisting endometriosis implants (Table 3). In these cases, additional samples for a thorough search for residual lesions of endometrial glands were not all successful. In 1925, Sampson [13] put forward the criteria for the diagnosis of malignant transformation of endometriosis: (a) benign and neoplastic endometrial tissue were shown in the tumor at the same time; (b) histology was compatible with the origin of the endometrium; (c) no other primary tumor sites were found. Besides, Scott proposed the fourth criterion in 1953 [14], that was, (d) the morphological manifestation of benign endometriosis adjacent to malignant tissue is a prerequisite for judging malignant tumors originating from endometriosis. In the current situation, the first three criteria of the reported cases have been met, but the fourth criteria may not be met, which may be the result of the complete replacement of normal tissue due to the proliferation of a large number of tumors.
Table 3
Pathological characteristics of lumps.
Tumor size(cm) | | Histology | Coexisting endometriosis |
Average | Range | CCC | CCC + EC | yes | no |
Case(n) | Percentage | Case(n) | Percentage | Case(n) | Percentage | Case(n) | Percentage |
10.0 ± 4.4 | (3,25) | 43 | 95.60% | 2 | 4.40% | 22 | 51% | 23 | 49% |
CCC: clear cell carcinoma, EC: endometrial cancer. |
Because some literature did not specify the pathological characteristics of lumps, the results of Pathological characteristics of lumps only counted the cases with specific results. |
Also, we classified the tumor into large masses and small masses by 10 cm in diameter. The correlation between mass size and clinical features was shown in Table 4. The results suggested that patients with abdominal masses larger than 10 cm had more symptoms of abdominal pain before admission than those with small masses, and complained more about the pain caused by palpable masses (P = 0.015). There was no significant difference in other characteristics (P > 0.05).
Table 4
The correlation between mass size and clinical features.
Clinical data | L-SE (8 cases) | S-SE (16 cases) | P |
Mean age (yr) (range) | 47.8 (38–57) | 46.8(38–60) | NS |
No. of gynecological sugeries (range) | 1.3 (1–3) | 1.5(1–3) | NS |
Delay from first surgery(y) | 14.8 ± 7.7 | 18.0 ± 8.3 | NS |
Onset of symptoms (m) | 66.5 ± 74.9 | 34.6 ± 39.9 | NS |
pain | 7 | 8 | 0.015 |
no pain | 1 | 8 |
L-SE: Large scar endometrioma(༞10cm); S-SE: Small scar endometrioma; NS: Not significant. |
NS: not significant. |
In terms of treatment, four patients received preoperative neoadjuvant chemotherapy (mainly platinum-based) [15–18], and one of them did not consider radical surgery because of the rapid progression of multiple metastases throughout the body [15]. Six patients were trying to treat vaginal estradiol medroxyprogesterone injection, leuprolide acetate, gestrinone, riptorelin, leiprim [10, 19–22], including our case, but the effect was not satisfactory. Surgery is the main treatment for the most of patients. The first-stage operation was performed based on extensive resection of the tumor and extensive abdominal tissue. Two cases were not treated surgically because of multiple metastases throughout the body and the rapid progression of the disease [15, 23]. Due to the extent of the fascia defect, 22 patients (47.8%) used mesh to reconstruct the abdominal wall (including our case]. Other regular operations include hysterectomy (27 cases, 60.0%) and/or salpingectomy (29 cases, 64.4%), and / or lymph node dissection (14 cases, 30.4%) and/or omental resection (15 cases, 33.3%). Although radical surgery is a major part of the treatment of all patients, our case did not performe hysterectomy and bilateral salpingo-oophorectomy plus pelvic lymphadenectomy, because no signs of the above organs were involved. Another reason is that we plan to perform postoperative radiotherapy for the patient after chemiotherapy. The follow-up adjuvant therapy is mainly chemotherapy, usually platinum and paclitaxel drugs. Twenty-nine cases (64.4%) received 1–9 cycles of adjuvant chemotherapy. However, due to the poor individual characteristics and compliance, adverse reactions, or partial reactions to the treatment, the treatment is interrupted and the efficacy of chemotherapy is difficult to evaluate. Only 16 patients (34.8%) received radiotherapy after surgery and chemotherapy (Table S1).
The above patients obtained the latest clinical follow-up information currently available, and eventually, 38 patients were able to know the outcome (Table S1). The follow-up time of the above patients ranged from 1 to 60 months, with an average of 15.0 months. 34.2% of women (13/38) died between 5 and 48 months after diagnosis. As shown in the Kaplan-Meier survival curve (Fig. 7), the median survival time is 48 months and the five-year survival rate is 35%. Seventeen patients were reported to relapse. The sites of recurrence were local (5 cases), lymphatic metastasis (3 cases, all inguinal lymph nodes) [15, 20, 24], distant metastasis (4 cases of liver, 2 cases of lung, 2 cases of brain, and 2 cases of bone). Univariate Cox regression model showed that the prognosis of patients with metastasis (17 cases, 17/37, Of the 38 cases with a specific follow-up period, one death was excluded because it was not known whether there was metastasis.) was worse than that without metastasis (20 cases, 20/37), as shown by Kaplan-Meier curve (Fig. 8), suggesting that metastasis invasion is a poor prognostic factor affecting patient survival. This is consistent with Taburiaux's study [25].
At present, there is no multicenter RCT in the study of abdominal CCC. However, we can put forward some general directions for future treatment and research, and emphasize several findings that seem to appear from our data: (a) abdominal wall scar CCC appears in relatively young women and is an invasive disease with a poor prognosis, with a 5-year survival rate of about 35%. (b) it is an iatrogenic disease of abdominal wall scar formation after gynecological surgery, and CS is the most common. (c) the progression of the disease was slow, and the average time of diagnosis was 17.9 years after the first gynecology and obstetrics operation. In this case, it often relapses to benign nodules of endometriosis, and sometimes patients undergo repeated surgery. (d) the size of the tumor was large at the time of diagnosis, and most of them needed extensive surgical repair of abdominal wall defects. (e) the use of progesterone or gonadotropin-releasing hormone analogs (GnRH-a) before surgical treatment is limited to the number of cases, and the effects are not clear.
The limitation of the review lies in the rarity of the disease, resulting in data collection based only on case reports, with heterologous information from different specialties. This leads to a lack of useful data, which limits statistical analysis. Although we have carried out a lot of statistical analysis, no clear results have been obtained. However, the results of this analysis can guide us to better understand the CCC risk factors and optimal treatment of abdominal wall endometriosis.