Preoperative prediction of malignancy and surgical treatment strategy in appendiceal tumors: multicenter review of 51 consecutive cases

A diagnostic and treatment strategy for appendiceal tumors (ATs) has not been established. We aimed to evaluate our treatment strategy in ATs, including laparoscopic surgery (LS), and to identify preoperative malignancy predictors. A total of 51 patients between 2011 and 2021 were retrospectively reviewed. Data, including tumor markers and imaging findings, were compared between carcinoma and non-carcinoma patients. Validity of planned operation was evaluated based on pathological diagnosis. Twenty-five patients were diagnosed with carcinoma, 13 with low-grade mucinous neoplasm, and 13 with other diseases. Symptoms were more commonly present in carcinoma patients than in non-carcinoma patients (68.0% vs. 23.1%, p = 0.001). Elevated CEA and CA19-9 were more frequently observed in carcinoma patients than in non-carcinoma patients (p < 0.01 and p = 0.04, respectively). Five carcinoma patients had malignancy on biopsy, compared with zero non-carcinoma patients. Significant differences were noted in the percentages of carcinoma and non-carcinoma patients with solid enhanced mass (41.7% vs. 0%, p < 0.001) and tumor wall irregularity (16.7% vs. 0%, p = 0.03) on imaging. Although the sensitivity was not high, the specificity and positive predictive value of these findings were 100%. Forty-two patients (82.4%) underwent LS as minimally invasive exploratory and/or radical operation, of whom 2 were converted to open surgery for invasion of adjacent organ. No patients had intraoperative complications or postoperative mortality. Clinical symptoms, elevated tumor markers, and worrisome features of solid enhanced mass and tumor wall irregularity on imaging can be malignancy predictors. For management of ATs, LS is feasible and useful for diagnosis and treatment.


Introduction
Appendiceal neoplasms are rare diseases, being identified in 0.9-2% of appendiceal specimens [1,2]. Preoperative histological diagnosis of appendiceal tumors is difficult because of the anatomical characteristics, and appendiceal neoplasms are sometimes diagnosed unexpectedly after emergency surgery for acute appendicitis. Appendiceal tumors have a wide variety of types. In the 5th edition of the World Health Organization (WHO) classification, these tumors are classified into the following categories: adenocarcinoma, mucinous neoplasm, serrated lesion/polyp, and neuroendocrine neoplasm (NEN) [3].
The most common malignancies of the appendix are mucinous and non-mucinous adenocarcinomas. Furthermore, 15.8% of mucinous and 43.9% of non-mucinous adenocarcinoma patients have lymph node metastasis (LNM) [4], and LNM is an independent factor for recurrence-free survival and overall survival in appendiceal carcinoma patients [5]. Although the clinicopathological and molecular characteristics of appendiceal carcinoma differ from those of colorectal cancer (CRC), patients with appendiceal carcinoma who underwent right-side colectomy (RC) with lymph node dissection (LND) had a better prognosis than those who underwent appendectomy alone, and therefore RC with LND is usually recommended as a radical surgery [6][7][8].
Mucinous neoplasms include low-grade appendiceal mucinous neoplasm (LAMN), high-grade appendiceal mucinous neoplasm (HAMN), and mucinous adenocarcinoma. Regardless of the histological grade, mucinous neoplasms can lead to mucocele formation and pseudomyxoma peritonei (PMP) [9]. For LAMN or HAMN, appendectomy alone with negative resection margin is considered sufficient, while mucinous adenocarcinoma usually requires RC with LND [10]. However, reaching a differential diagnosis between LAMN/HAMN and carcinoma is often difficult preoperatively, and thus the surgical treatment strategy in mucinous neoplasms remains controversial. To reach a decision for the surgical approach, detection of these tumors on preoperative imaging is important. Computed tomography (CT) can help to confirm an appendiceal tumor and may suggest a specific diagnosis [11]. However, CT cannot provide a histological diagnosis, unlike colonoscopy in CRC, and thus a preoperative definite diagnosis is difficult unless the tumor invades the appendiceal orifice and/or cecum. It is also difficult to confirm the preoperative diagnosis of other rare tumors, including NEN, lymphoma, and benign disease, because these tumors have fewer specific imaging findings. Only a few studies have focused on the preoperative findings, diagnosis, and treatment strategy in appendiceal tumors to date [12].
Recently, the number of procedures involving laparoscopic surgery (LS) has been increasing in a variety of surgical fields [13]. Several studies reported that LS can be used for safe and radical resection of CRC and appendiceal tumors [14]. LS may have advantages of not only reduced invasiveness but also utility as an intraoperative diagnostic tool for cases without a definite preoperative diagnosis.
The aim of this study was to determine malignancy predictors using clinical data and preoperative imaging findings in patients with appendiceal tumors and to evaluate the validity of performed operations based on the preoperative and postoperative diagnosis. We also discuss the usefulness of LS in appendiceal tumors and review our surgical treatment strategy in appendiceal tumors.

Methods
The present study was approved by the ethics committees of Hamanomachi Hospital and Kyushu University Hospital (approval numbers 2020-48 and 2020-503, respectively) and was conducted according to the ethical guidelines of the Helsinki Declaration. The medical records of 51 consecutive patients who were preoperatively and/or histologically diagnosed with appendiceal tumors at the Department of Surgery, Hamanomachi Hospital, and the Department of Surgery and Oncology, Kyushu University Hospital, between 2011 and 2021, were retrospectively reviewed. Data including demographic characteristics, clinical symptoms, preoperative imaging findings, preoperative and postoperative diagnosis, operative findings and procedures, and postoperative outcomes were collected.
If available, serum levels of carcinoembryonic antigen (CEA) (normal limit, < 5.0 ng/mL at Hamanomachi Hospital; < 3.2 ng/mL at Kyushu University Hospital) and carbohydrate antigen 19-9 (CA19-9) (normal limit, < 37 ng/ mL) were reviewed. For preoperative assessment, enhanced CT and/or magnetic resonance imaging (MRI) findings were evaluated in all patients. In cystic tumor patients, the maximum tumor diameter was determined based on preoperative CT/MRI. Acute appendicitis patients who were postoperatively proven to have appendiceal carcinoma were included, while patients who underwent appendectomy for acute appendicitis without a neoplastic lesion were excluded.
Based on the preoperative imaging findings and/or the tumor biopsy findings during colonoscopy, an appropriate operative procedure was selected in each patient (Fig. 1). The first choice for the operative procedure was LS. If LS was not applicable, open surgery was performed. For a definite malignant tumor or likely malignant tumor, laparoscopic RC (including right-hemicolectomy and ileocecal resection) (LS-RC) with LND was planned. For a likely nonmalignant tumor or apparent benign tumor, laparoscopic appendectomy or cecectomy (LS-A/C) without LND was performed. RC was also selected in nonmalignant cases to obtain the appropriate surgical margin during surgery. Intraoperative frozen pathology was evaluated during LS-A/C as needed, and if the pathological diagnosis was malignant tumor, LS-RC with LND was subsequently performed in the same operation. If the appendectomy specimen for appendicitis was found to contain advanced carcinoma, additional RC with LND was performed. For patients with suspicion of PMP, exploratory laparoscopy with or without appendectomy was planned first, and further treatment was considered.
Based on the 2019 WHO Classification, the postoperative diagnosis in each patient was categorized into the following cohorts. Carcinoma included appendiceal cancer (adenocarcinoma, mucinous adenocarcinoma, and goblet cell carcinoma) and cecal cancer. Cecal cancer around the orifice of the appendix, which had been preoperatively diagnosed as appendiceal tumor, was also included because it obstructed the appendiceal lumen and made the appendix swollen. Other tumor lesions were LAMN and neuroendocrine tumor (NET), and the remaining rare conditions, such as malignant lymphoma (ML), adenoma, sessile serrated adenoma/ polyp (SSA/P), hyperplastic polyp, chronic appendicitis, and diverticulitis were categorized as others. Patients with PMP originating from appendiceal tumors were included.
Comparisons between carcinoma and non-carcinoma patients were performed using the chi-square test or Mann-Whitney U-test as appropriate. A probability value of < 0.05 was considered statistically significant. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of each preoperative finding for the diagnosis of malignancy were calculated. All statistical analyses were conducted using JMP Pro version 15.1.0 (SAS Institute Inc., Tokyo, Japan).
Twenty-three patients had symptoms: 12 had abdominal pain, 2 had back pain, 2 had melena, 1 had fever, 2 had palpable abdominal mass, and 4 were diagnosed with acute appendicitis at the initial operation. Seventeen of the 25 carcinoma patients had at least one symptom, as did 2 of the 13 LAMN patients, and 4 of the 10 patients with other diseases. None of the 3 NET patients had symptoms. When the numbers of patients with clinical symptoms were compared, symptoms were significantly more common in carcinoma patients than in non-carcinoma patients (p = 0.001; Table 1).

Surgical management flow
As shown in Fig. 1, the surgical procedure was selected based on the preoperative tumor markers and imaging findings, or definite pathological diagnosis obtained. Thirtythree patients who were preoperatively diagnosed with  Table 1) who was preoperatively diagnosed with mucinous cystic tumor underwent LS-RC to obtain the surgical margin. Among the 33 patients, the pathological diagnosis was carcinoma in 22 (appendiceal carcinoma in 21 and cecal carcinoma in 1), LAMN in 9, ML in 1, and diverticulitis in 1. In the ML patient, there was a huge mass around the appendix with cecum and sigmoid colon invasion, causing abdominal pain. Although ML was suspected, upfront resection was performed to avoid the risk of tumor perforation during chemotherapy. In 3 patients who underwent appendectomy with a diagnosis of acute appendicitis and were proven to have advanced appendiceal carcinoma, RC with LND was subsequently performed as a curative operation (cases H12, H13, and K39; Supplemental Table 2).
LS-A/C was performed for likely non-carcinoma lesions in 12 patients, comprising 4 with LAMN, 3 with chronic appendicitis, 2 with adenoma, 1 with SSA/P, 1 with hyperplastic polyp, and 1 with appendiceal endometriosis. In 4 of these 12 patients, intraoperative frozen pathology was evaluated to determine whether the appendectomy specimens contained malignant tissue, and none of them required additional resection. In 3 patients who underwent LS/open-A with a diagnosis of chronic appendicitis, the pathological diagnosis was NET-G1, which did not require additional treatment. Three patients with likely PMP underwent exploratory surgery and received a definite diagnosis. Of these 3 PMP patients, 1 underwent cytoreductive surgery (CRS), 1 was planned to undergo CRS with hyperthermic intraperitoneal chemotherapy (HIPEC) at another institution but only underwent an exploratory laparotomy, and 1 refused additional treatment.
Among the patients with LAMN, no patients had LNM, distant metastasis, or PMP. The median surveillance period was 26 months (range: 5-75 months). None of the patients had recurrence of mucinous neoplasms during the surveillance period.

Analysis of preoperative serum levels of tumor markers
Among the patients whose serum CEA levels were measured preoperatively, 13 of 21 carcinoma patients, 4 of 11 LAMN patients, and 1 chronic appendicitis patient with concomitant gastric cancer had elevated CEA. There was a significant difference between the numbers of carcinoma and non-carcinoma patients with elevated serum CEA (p < 0.01;  Table 1). For elevated CEA as a malignancy predictor, the odds ratio was 5.55 (95% confidence interval, 1.50-20.28), the PPV was 0.72, and the NPV was 0.68. Similarly, among the patients whose serum CA19-9 levels were measured preoperatively, 5 of 17 carcinoma patients showed elevated CA19-9, while only 1 LAMN patient among 21 non-carcinoma patients had elevated CA19-9. The number of carcinoma patients with elevated serum CA19-9 was significantly higher than that of non-carcinoma patients (p = 0.04; Table 1). For elevated CA19-9 as a malignancy predictor, the odds ratio was 8.33 (95% confidence interval, 1.10-59.02), the PPV was 0.83, and the NPV was 0.63.

Analysis of preoperative imaging findings
Forty patients, including 18 carcinoma patients, underwent preoperative colonoscopy, and malignant tumor was suspected in 8 patients. The other patients showed normal findings or submucosal tumor-like elevation. Of the 8 patients, 5 were diagnosed with adenocarcinoma and 1 with ML by biopsy, consistent with the postoperative diagnosis. Among the patients who underwent colonoscopy, the sensitivity of positive biopsy was not high (27.8%), but the specificity and PPV were 100% ( Table 2).
The preoperative imaging findings in 24 carcinoma patients (imaging was unavailable in 1 patient) are shown in Table 2. Among these patients, 10 had solid enhanced lesion, 3 had cystic tumor with solid component, 4 had tumor wall irregularity, 3 had enhanced tumor wall thickness, 2 had calcification, and 3 had apparent PMP findings. There were significant differences between the numbers of carcinoma and non-carcinoma patients having solid enhanced mass (p < 0.001) and tumor wall irregularity (p = 0.03) (Fig. 2a-d). Solid enhanced mass showed the highest sensitivity (41.7%), while the sensitivity of the other features was not high. However, the specificity and PPV of solidenhanced mass and tumor wall irregularity were both 100%, indicating that these findings are good malignancy predictors (Table 2). Meanwhile, the specificity and PPV of cyst with solid component were 96.2 and 75%, respectively.
Regarding the tumor size of mucinous neoplasms, the median mucocele size of LAMN was 50 mm (range: 20-120 mm), and the median size of mucinous adenocarcinoma was 55 mm (range: 21-100 mm). There were no significant differences in the tumor sizes between the carcinoma and non-carcinoma patients.

Laparoscopic surgery
LS was selected in 42 of the 51 patients (82.4%), of whom 1 ML patient and 1 carcinoma patient required conversion to open surgery for invasion of adjacent organ by the tumor (case K34 in Supplemental Table 3 and case K46

Discussion
In this retrospective multicenter review, the authors identified worrisome features for malignancy of appendiceal tumors. These features included clinical symptoms and elevated serum CEA and/or CA19-9 as well as solid enhanced mass and/or tumor wall irregularity on CT or MRI. The majority (96%) of the patients in our carcinoma cohort had at least one of these findings. Thus, if colonoscopy is unavailable or fails to reveal carcinoma, RC with LND is recommended as a curative operation for treatment of appendiceal tumors in patients with these findings. An algorithm for the management of appendiceal tumors based on the present findings is shown in Fig. 3. Although some researchers reported a retrospective case series of 13 appendiceal cancer patients in which carcinoma was not suspected preoperatively [7] and other researchers demonstrated that surrounding soft tissue thickening and tumor wall irregularity were nonspecific findings because they could be caused by inflammation as well as malignancy [11,12], we consider that the relatively high PPV of each feature may help us to select the proper treatment.
Among the 18 carcinoma patients who underwent colonoscopy followed by biopsy in our cohort, carcinoma was suspected in 7 (38.9%) and the biopsy was positive for carcinoma in 5 (27.8%). Trivedi et al. [15] reported that a malignant tumor diagnosis was made in only 2 of 64 (3.1%) appendiceal carcinoma patients showing PMP by biopsy during preoperative colonoscopy, and their positivity rate for cancer was lower than that in the present study. This discrepancy may be due to the biological differences between mucinous adenocarcinoma, in which the tumor cells tend to invade the appendiceal wall with abundant mucin, and tubular adenocarcinoma, in which the cells tend to form a solid mass and sometimes spread over the lumen into the cecum. Although its diagnostic sensitivity is not sufficiently high, colonoscopy may reveal indirect findings, such as cecal or ileal mucosal changes, and may indicate the extent of the disease. Thus, we recommend total colonoscopy in all patients with appendiceal tumors if applicable, especially in patients with the worrisome features of carcinoma.
When considering appendiceal cystic tumor cases only, it is important to distinguish mucinous adenocarcinoma from LAMN and other benign lesions because LNM or distant metastasis is significantly less common in LAMN patients than in mucinous adenocarcinoma patients [4,5,16]. In our LAMN cohort containing 4 patients who underwent LS-A/C, no patients had LNM or developed postoperative recurrence during the surveillance period. Because patients Fig. 2 Representative imaging findings as worrisome features on abdominal contrast-enhanced computed tomography. a A typical solid enhanced mass (white arrow), which was diagnosed as non-mucinous adenocarcinoma. b A non-enhanced solid mass lesion at the peri-appendix (white arrowheads). The pathological diagnosis was appendiceal endometriosis. c A cystic mass with enhanced tumor wall irregularity (white arrowhead). The pathological diagnosis was mucinous adenocarcinoma. d A cystic and dilated appendix without tumor wall irregularity (black arrow). The pathological diagnosis was low-grade appendiceal mucinous neoplasm with LAMN were shown to have a good prognosis and RC with LND did not contribute to increased disease-specific or overall survival, extended resection in LAMN may not be necessary [17,18]. While all 9 mucinous adenocarcinoma patients in our cohort showed at least one of the worrisome features, there was no difference between mucinous carcinoma and LAMN in tumor size, with the mean diameter being as large as 5 cm. Therefore, these features can assist surgeons to choose appropriate operative procedures for patients with large cystic tumors.
Regardless of the histological grade, mucinous neoplasm can lead to PMP. In our study, all 3 PMP patients were histologically confirmed as mucinous carcinoma, and no LAMN patients were accompanied by PMP. Of the 3 PMP patients, 1 underwent CRS alone, and 1 was planned to undergo CRS with HIPEC at another institution but it could not be performed. Combination therapy of CRS and HIPEC has been recognized as a standard protocol for PMP, although it was reported to increase the 60-day surgical morbidity [19,20]. Further prospective studies with larger samples of PMP patients are needed to determine the oncological efficacy.
The frequency of epithelial tumors has been increasing over the past few decades, according to the Surveillance, Epidemiology, and End Results (SEER) database [1]. In our study, the most frequent neoplasm was appendiceal cancer, similar to the SEER database, and these neoplasms were often discovered unexpectedly in resected specimens. Actually, 3 patients were diagnosed with adenocarcinoma after undergoing appendectomy for acute appendicitis.
Although it is usually difficult to identify a malignant tumor in acute appendicitis patients, malignant tumors in patients with appendicitis seem to be associated with increasing age [21]. The ages of our 3 patients were 46, 57, and 74 years, and were higher than the peak incidence age for typical acute appendicitis [22]. Furthermore, Sugimoto et al. [23] demonstrated that patients with complicated appendicitis had a higher incidence of appendiceal tumors than patients with uncomplicated appendicitis. In our cohort, 2 of the 3 patients had complicated appendicitis with abscess formation. For the management of acute appendicitis, especially in patients with older age or complicated appendicitis, surgeons should pay attention to Fig. 3 Algorithm for the management of appendiceal tumors. If an appendiceal tumor is suspected, the patient's symptoms, serum levels of tumor markers, and imaging findings, including colonoscopy should be evaluated. Presence of any one of the worrisome features is suggestive of a malignant appendiceal tumor, and thus, right-side colectomy with lymph node dissection is recommended. CEA, carcinoembryonic antigen; CA19-9, carbohydrate antigen 19-9; TCS, total colonoscopy; PMP, pseudomyxoma peritonei; RC, rightside colectomy; LND, lymph node dissection; HIPEC, hyperthermic intraperitoneal chemotherapy the possible presence of malignancy and should confirm the levels of CEA and CA19-9 preoperatively. It is also important to avoid spilling the contents of the appendix during surgery, based on the possible existence of malignant cells [24].
Regarding other neoplasms, NEN was diagnosed in 0.1-0.3% of patients after appendectomy in a previous study [25]. In the present study, 3 patients with NET-G1 were incidentally diagnosed after resection, and none of these patients showed any abnormalities of the appendix on retrospective assessment of their preoperative imaging findings. The majority of appendiceal NEN cases are < 1 cm in size and confined to the distal appendix, which may render a preoperative diagnosis difficult [11,25]. Regarding ML, these cases are also incidentally diagnosed as causes of acute appendicitis [26]. In the present study, there was only one ML patient, and the symptom, in this case, was acute appendicitis. The published articles on imaging features of appendiceal lymphoma are mainly limited to case reports, and thus the specific imaging features remain unknown [26]. Although we selected a treatment strategy for these neoplasms based on the corresponding guideline, preoperative histological diagnosis of such neoplasms is difficult unless the tumors have a relatively large size and invade the cecum.
Recently, the number of cases treated by LS has been increasing, and decreased rates of intraoperative and postoperative complications have been reported [13]. In our study, 42 patients (82.4%) were planned to undergo LS, and only 2 of these patients (4.8%) were converted to open surgery for invasion of adjacent organs. Kim et al. [14] compared the perioperative short-term outcomes between LS and open surgery cases and showed the safety and feasibility of LS in appendiceal mucocele. Inoue et al. [27] focused on the longterm outcomes of LS in appendiceal tumors and showed that LS was comparable to open surgery. In addition to the superior short-term and long-term outcomes, another advantage of LS is the possibility for diagnostic exploration of the appendix itself and entire abdomen to investigate peritoneal dissemination under magnified vision. Moreover, LS-A/C is relatively easy to perform and allows acquisition of a specimen for intraoperative pathological diagnosis. Diagnostic LS followed by A/C would be a better option, especially in older and/or vulnerable patients with appendiceal tumors for which cancer could not be ruled out.
The present study has some limitations. It was a retrospective study, and the total number of cases was relatively small. Consequently, the statistical power was weak, and it was relatively difficult to detect definite malignant features in preoperative findings. In addition, the median postoperative surveillance periods varied, and the long-term outcomes were unknown, especially in LAMN patients. Further investigations are necessary to establish a preoperative diagnosis and treatment strategy in appendiceal tumors.
In conclusion, although definite preoperative diagnosis of an appendiceal tumor is difficult, clinical symptoms, preoperative tumor markers, and some imaging findings can be worrisome features for malignancy. The possible presence of malignancy should be kept in mind in patients with older age or complicated acute appendicitis. For the clinical management of appendiceal tumor patients, LS is feasible and useful for intraoperative diagnosis and surgical treatment.