Great Mimics in Oncology: A Retrospective Study from a Tertiary Care Centre of Eastern India

Abstract

A number of benign diseases can masquerade as malignancy leading to unnecessary treatment. Vice versa, many benign-looking tumours when operated turns out to be malignant. While the latter necessitates extra surgery for oncological clearance, the former directly harms the patient impacting their lives seriously. Data pertaining to such “misdiagnosis” is scarce and there is an urgent need to document such cases to prevent public harm. We carried out a retrospective study to identify characteristic of such cases which were actually benign but operated upon with a diagnosis of malignancy. This is a retrospective study done at the Department of Surgical Oncology, Institute of Post Graduate Medical Education & Research (I.P.G.M.E&R). Databases from January 2022 to August 2023 were searched for patients who were initially diagnosed as cases of malignancy but later turned out to be benign. Demographic and clinicopathological data were retrieved and analysed. Out of 345 major cases, 18 cases were misdiagnosed as cancer. Three cases mimicked breast lump, two cases misdiagnosed as lymphoma, and one case each diagnosed as primary peritoneal carcinoma, carcinoma ovary, carcinoma gallbladder, and soft tissue tumour. Two cases turned out to be tuberculosis (TB), and one case was rare Castleman disease, while an unusual diagnosis of Ig4 disease was made on histopathology. Although mortality was zero, one patient had perioperative morbidity in the form of bleeding, post-op infection, and prolonged hospital stay while another patient received intraoperative brachytherapy unnecessarily. Out of 18 cases, ten cases had a preoperative cytology report suggestive of neoplasm, in three cases the biopsy/fine needle aspiration cytology (FNAC) report was inconclusive, while five patients were diagnosed based solely on clinical and radiologic findings due to an inconclusive cytology report. A benign entity can mimic cancer almost anywhere in the body. Due to close clinical, radiologic, and cytological findings, such situations are not uncommon in day to day practice. High degree of suspicion, good interdisciplinary communication, and review of slides by an experienced cytopathologist can help prevent such misdiagnosis to a good extent.

Introduction

The rate of misdiagnosis in medical practice is a topic one seldom talks about. The exact definition of “misdiagnosis” remains unknown; hence, exact incidence rates are not available for comparison. In general, it is said that most people face one diagnostic error in their lifetime. One study quotes that the rates of underdiagnosis or misdiagnosis can be as high as 44%[1].

.Cancer-specific misdiagnosis whereby benign diseases are misinterpreted as malignancy or vice versa is a serious cause of concern and causes tangible harm to the patient. The harm rates associated with such cancer-related misdiagnosis remain unknown. A number of benign diseases or entities notably of infectious, inflammatory, traumatic, or immunologic origin have been reported at various points of time to mimic malignancy due to similar clinical, radiological, and histological appearances. Studies documenting such cases are rare in Indian literature. The characteristics of such cases and their impact on patient management are unknown. Available studies are limited to case reports or case series involving only single disease site or subsites. We carried out this study to characterise clinicodemographic factors associated with benign cases masquerading as malignancy and their impact on patient outcomes.

Materials and Methods

Objectives

To study the demographic and clinicopathologic characteristics of benign cases being diagnosed as malignancy.

Inclusion Criteria

All indoor patients registered under the Department of Surgical Oncology between February 2019 and January 2021.

Exclusion Criteria

1. OPD patients were excluded from the study

2. Those who did not consent for study (including all patients who did not agree to use and share their personal data for analysis even after proper counselling)

Type of Study

Retrospective observational hospital–based study.

Site

Department of Surgical Oncology, I.P.G.M.E&R, Kolkata.

Method

Patients were admitted after consultation in OPD where on average 50 new and old cases are seen twice weekly. The case records of all patients who were admitted during the study duration with intent for surgery were scrutinized. The details of cases where there was a change in final diagnosis from malignancy to benign entity, based on final histopathological reports, were selected and included for analysis. Those data were tabulated and entered into an Excel sheet for analysis.

Results

Eighteen out of 345 (4%) cases were misdiagnosed as cancer (Table 1). Most were females (12), while the rest were males (6). Most belonged to the middle-aged group and most came from villages. The median duration of symptoms was 4 months. No history of addiction to tobacco or alcohol was found in the entire cohort. Seventeen cases were referred to us from various hospitals while one case was on old follow-up case of carcinoma parotid. In 15 out of 18 cases, there was concordance between clinical, radiological, and cytological findings in favour of malignant diagnosis. In three cases (all breast cases), although clinically and radiologically they mimicked breast lumps, cytology was inconclusive. Out of the total number of cases, five were infective in origin (tubercular) and three cases were related to an immune-mediated disease, while three were inflammatory in nature while the rest we grouped as “ miscellaneous”.

Table 1.

Demographic characteristics of study sample

Age	25–50 (in years) 17	> 50 (in years) 1
Sex	Male 9	Female 6
Residence	Urban 6	Rural 12
Presentation	Symptomatic 16	Incidental 2
Addiction	Smoking 0	Alcohol 0

We divided 15 cases where surgery was done with presumed diagnosis of cancer into two categories for better discussion (Fig. 1):

Fig. 1.

Schematic flowchart of overall treatment done

Group A: Treatment that could be justified as surgery would have to be performed any way, based on patient symptoms, complaints, or for diagnostic confirmation (seven cases).

Group B: Treatment could be justified but not the radicality of surgery (eight cases). In other words, patients were overtreated, i.e. the extent of surgery could have been less when final diagnosis was taken into consideration.

This group (Table 2) had one case of pelvic endometriosis, presenting with complex abdominopelvic mass. All of them presented with abdominal distention, unresolving fatigue, and loss of body weight along with raised CA-125 levels. All of them underwent staging laparotomy with frozen section control. A frozen section reported to be suspicious would commit us to radical resection needed for ovarian cancer surgery.

Table 2.

List of all cases analysed and treated

Provisional diagnosis	Final diagnosis
Suspected breast lump	Granulomatous mastitis
Breast lump with supraclavicular lymph node	Tuberculosis
Neck mass (left)	Rosai-Dorfman
Neck mass (right)	Castleman DISEASE
Neck mass with regional lymphadenopathy (lymphoma)	Tuberculosis
Retroperitoneal mass	Ig4-related disease
Gastric mass/GIST	Old perforation with inflammatory mass
Advanced ovarian carcinoma	Peritoneal tuberculosis
Suspected ovarian ca/pelvic mass	Endometriosis
Gall bladder carcinoma/mass	Xanthogranulomatous cholecystitis
Gastric carcinoma with outlet obstruction	Peptic stricture
Calf mass/sarcoma	Myositis ossificans
Ca thyroid (incidentaloma)	Thyroiditis

There was one case of abdominal ascites with radiologic evidence of omental caking and history of repeated paracentesis. Ascitic fluid cytology and image-guided biopsy were repeatedly reported negative. She underwent laparotomy with presumed diagnosis of advanced ovarian cancer. However, she was found to be unfit for optimal cytoreduction, and a biopsy in the form of omentectomy was done. Later on, the biopsy was consistent with tuberculosis.

One patient had complaints of abdominal tenderness, anorexia, and loss of weight. Although upper GI endoscopy was negative, radiologic scans were suggestive of an intra-abdominal mass, of probably gastric origin. Fine needle aspiration cytology (FNAC) was suggestive of gastrointestinal stromal tumour (GIST). However, an exploratory laparotomy revealed the presence of old peptic perforation and only intra-abdominal abscess. In another similar case, a patient was referred to us with diagnosis of retroperitoneal sarcoma. On exploratory laparotomy, he had an unresectable mass. Biopsy proved it to be a rare case of an Ig4-related disease.

In one biopsy-proven case of gastric adenocarcinoma presenting with gastric outlet obstruction, the final histopathological examination (HPE) revealed peptic stricture, to our surprise. Another patient presented with dyspepsia, and on radiologic workup, his provisional diagnosis was gallbladder carcinoma. He underwent radical cholecystectomy with portal lymph node dissection. His final HPE was xanthogranulomatous cholecystitis.

Among the two neck masses suspected to have lymphoma, one case turned out to be rare Rosai-Dorfman disease and the other Castleman disease respectively.

The other two cases were those of thyroid incidentaloma, where we performed total thyroidectomy (based on FDG-PET positivity and FNAC suggestive of papillary carcinoma thyroid) and calf soft tissue mass which turned out to be a case of myositis ossificans.

Although there was no mortality in the entire series, one patient in group B had postoperative bleeding along with prolonged ICU stay while another patient suffered from delayed complications of intraoperative radiotherapy. At 1-year follow-up, all patients are healthy. On a further detailed analysis of group B (Table 3), it was found that out of eight cases, in two cases (myositis ossificans and total thyroidectomy), the error could have been avoided if the slides were reviewed and appropriate clinical history considered while making a cancer diagnosis. In the other six cases, there was no way of avoiding the surgery due to their presentation.

Table 3.

Analysis of 8 cases where definite overtreatment was done

Diagnosis	Presentation	Final diagnosis
Soft tissue sarcoma (calf)	Swelling in calf for 2 years	Myositis ossificans
Thyroid nodule (biopsy-proven PTC)	Incidentaloma on PET Scan	Thyroiditis
Gastric mass/GIST (FNAC-spindle cell neoplasm)	Gastric outlet obstruction	Old perforative peritonitis
Ovarian/pelvic mass with raised makers (2 cases)/intra op frozen suspicious of malignancy	Pelvic lump	Endometriosis
Gallbladder mass	Weight loss and lump in abdomen	Xanthogranulomatous cholecystitis
Gastric carcinoma/peptic stricture (biopsy –adenocarcinoma, repeat biopsy –high-grade dysplasia)	Gastric outlet obstruction	Peptic stricture
Large neck mass (FNAC-spindle cell neoplasm), core cut biopsy—inconclusive	Large neck lump	Rosai-Dorfman disease

Discussion

Cancer mimicry or benign cases masquerading malignancy or vice versa are important both to clinicians and the patient. For the treating oncologist, it can be a source of clinical dilemma and utter frustration. Sometimes, such cases can lead to medicolegal litigation as well[2]. For the patient, a wrong cancer diagnosis can lead to unnecessary investigations and invasive interventions in the form of unwanted surgeries and all their associated complications. In addition to the costs incurred, the psychological consequences of a misdiagnosis in the form of cancer are immeasurable. Yet, literature on such an important issue is sparse and needs further research. According to our study, the rate of such misdiagnosis is 5.2% which is much lesser than the quoted rates in literature (10–15%)[3]. However, the rates may be higher as a large number of cases were not documented. The number of benign diseases or entities which can mimic malignancy is endless; however, it is important to be familiar with the most common ones. Infection, inflammation, and immune aetiologies were the three most common reasons for diagnostic error in our subset. Tuberculosis was the most common infective aetiology, and it is well known to mimic cases of lymphoma (presenting with lymphadenopathy), lung cancer (present as nodule, masses, parenchymal infiltrates, or effusions), ovarian cancer (often present with ascites, omental caking, tubo-ovarian mass, and raised CA-125), and breast cancer (lumps with lymphadenopathy), and needs to be considered when making a cancer diagnosis[4, 5]. Due to high prevalence and myriad presentations, tuberculosis should always be in the list of differential diagnosis as the rate of misdiagnosis associated with it can be as high as 44%[6]. Among the benign entities of inflammatory origin, idiopathic granulomatous mastitis deserves special mention as it frequently mimics breast lumps both clinically and radiologically leading to unnecessary oncology referrals[7]. We had three such cases which resolved with pulse-dose steroid therapy and limited debridement. Pelvic endometriosis, a hormone-dependent disease, was the common cause of error in three cases of carcinoma ovary taken for curative surgery. The diagnosis of endometriosis or endometriotic cysts often occurs as a surprise at laparotomy similar to what was seen in our study[8]. They closely masquerade ovarian cancer[9] and often present with complex cysts with raised tumour markers. Such cysts are often enhanced on computed tomography scans, and they are difficult to differentiate from malignancy even on MRI scans.

Distinguishing between advanced ovarian carcinoma and tuberculosis can be impossible as Mantoux test, sputum for AFB, and even ADA levels may be normal in rare cases and can easily lead to a diagnostic error as shown frequently by others in literature[10]. This happened in one such case in our series.

One case suspected to have GIST was finally found to be a simple case of old sealed perforative peritonitis, and similarly, in another case, where a patient underwent surgery for biopsy-proven gastric carcinoma had peptic stricture and metaplastic changes in final histopathology. We could not find any specific reference for this case in literature. Two cases were referred to us with neck swellings with provisional diagnosis of lymphoma. They were later found to be Castleman disease[11] which is well known for its ability to masquerade malignancy at any anatomical site.

There was one case of IG4-related unresectable intra-abdominal disease in our series. This rare entity is a known mimicker of malignancy, well reported in literature[12].

Xanthogranulomatous cholecystitis is a rare benign disease with close overlap with features of gallbladder carcinoma. Deng et al.[13] in 2015 in their large series of 42 such cases concluded that there is no practical way of differentiating between gallbladder cancer and xanthogranulomatous cholecystitis preoperatively as imaging, laboratory, and intraoperative frozen assessments can all be erroneous! We had a similar experience with one such case in our series.

One case presumed to have soft tissue sarcoma (based on FNAC, biopsy, and radiology) underwent sarcoma excision but was finally diagnosed to have myositis ossificans on histopathological examination. On literature review, we found several such cases[14] where even biopsy could be misleading if done too early, during evolution of myositis ossificans leading to inappropriate treatment.

Biopsy or cytological proof is prerequisite for most cancer treatments. However, the same may not be available due to various reasons. Difficulty in obtaining tumour tissue due to location of the tumour (for example proximity to major vessels, deep-seated tumours), lack of technical expertise, or inability to sample or target the tumour-bearing areas correctly are some of the reasons which lead to inconclusive results.

In some cancers like gallbladder cancer, testicular cancer, and ovarian neoplasms, needle biopsy is contraindicated in fear of tumour seeding and treatment is based primarily on information based on imaging finding, tumour marker level, and clinical findings, leading to inevitable chances of misdiagnosis. In our study, we had three such cases (one case of suspected gallbladder cancer and another ovarian mass) where we had to undertake surgical resection without cytological proof. Frozen section may be useful in such cases which has accuracy between 91.5 and 97.4%[15]. The diagnostic accuracy is different for different tumours. One study reported a diagnostic accuracy of 87.5% for ovarian neoplasms and reported less sensitivity for mucinous tumours[16]. Although reliable, frozen section results may be discordant with the final histopathological results as they are prone to technical, sampling, and interpretation errors.

Some tumours which tend to be very small like pancreatic neuroendocrine tumours are often too small to locate and are often not amenable to biopsy or FNAC. While FNAC is rapid and technically easy to perform, it is less ideal then core needle biopsy which is currently the gold standard of tissue acquisition and has higher specificity and sensitivity for most cancers and cancer subsites. However, core needle biopsies have lower diagnostic accuracy rates for sarcomas and abdominal wall tumours when compared with bony tumours[17]. Tumour heterogeneity and sampling errors are potential pitfalls of needle core biopsies, which can lead to false negative and inconclusive results as seen in our study.

Conclusion

In our series, infective and inflammatory causes were the commonest entities that mimicked cancer and led to misdiagnosis and overtreatment. Although in most cases frozen section, slide reviews, image-guided core cut biopsies, etc., when judiciously used in the appropriate clinical context can reduce the error in decision making, some errors are inevitable. Knowledge of common benign entities masquerading as cancer is essential to inform patient attendants about unexpected outcomes in the final pathological examination.

Limitation

Our study is limited by the retrospective nature of disease and small sample size. In few cases, slides and paraffin blocks of specimen were of suboptimal qualities which hampered review by our pathologists.

Data Availability

All data are freely available in this study.

Declarations

Conflict of Interest

The authors declare no competing interests