Brain Abscess Mimicking Brain Tumors: A Systematic Review of Individual Patient's Data

Abstract

Objectives Brain abscess is a worrisome condition with a 1-year mortality rate of 21% and a 32% rate of new-onset epilepsy. Brain magnetic resonance imaging (MRI) is strongly recommended as a screening modality with contrast-enhanced T1-weighted images, diffusion-weighted imaging (DWI), and attenuated diffusion coefficient. However, there is a 10% rate of false negative, which could potentially impact management and prognosis. Our systematic review aims at identifying risk factors for false negative.

Materials and Methods A database search of our institutions plus a systematic literature review was conducted using MEDLINE/PubMed, including studies of brain abscesses misdiagnosed as brain tumors. Data on patient demographics, clinical presentations, imaging findings, pathogens, treatments, and outcomes were extracted and analyzed. We present a case of a 59-year-old male with HIV, who developed a brain abscess misdiagnosed as a tumor. Initial symptoms included left-side weakness and weight loss. Imaging showed a ring-enhancing lesion in the right thalamus. The abscess was caused by T. gondii, and the patient was treated with sulfadiazine, pyrimethamine, ceftriaxone, and metronidazole, achieving a GOS-E score of 8 at 1 year.

Results The review included 14 studies, with 1 additional illustrative case, encompassing a total of 15 cases. Patients ranged from 39 to 77 years, with a mean age of 59 years. Comorbidities included human immunodeficiency virus (HIV), glioblastoma, breast cancer, arthritis, gastric cancer, and nephrotic syndrome. Common symptoms were hemiparesis, generalized seizures, headache, and confusion. Imaging often revealed ring-enhancing lesions with restricted diffusion on DWI. Lesions were located in various brain regions. Pathogens identified included 40% Nocardia species, Toxoplasma gondii, Mycobacterium tuberculosis, Aggregatibacter aphrophilus, Rickettsia typhi, Arcanobacterium haemolyticum, Aspergillus terreus, and Providencia rettgeri. Treatments involved antibiotics and, in some cases, surgical intervention. Outcomes measured by the Glasgow Outcome Scale-Extended (GOS-E) at 1 year indicated good recovery in most cases.

Conclusion Despite the high sensitivity and specificity of brain MRI in diagnosing brain abscesses, the standard protocol used for the past two decades still results in a 10% false-negative rate. Such inaccuracies can significantly impact the patient's management, potentially delaying antibiotic therapy and impacting the surgical planning, hence affecting the outcome. Immunocompromised patients are particularly vulnerable to misdiagnoses of brain abscesses as brain tumors. To improve diagnostic accuracy, new imaging techniques and computational tools are currently under investigation.

Keywords

brain abscess - systematic review - case reports - mimicking - toxoplasmosis - nocardia - tumor - MRI

Introduction

Brain abscesses (BAs) start of as focal infections of brain parenchyma, known as cerebritis, which develops into a collection of pus surrounded by a capsule. Pyogenic (bacterial) abscesses stand as the most prevalent ones. BA has an annual incidence rate of 0.9 per 100,000 persons,[1] and a 1-year mortality of 21%.[2] This condition not only affects mortality but also significantly affects the quality of life, evidenced by a 32% incidence of new-onset epilepsy for instance.[2] BAs typically present in magnetic resonance imaging (MRI) scans as ring-enhancing lesions on T1-weighted images with contrast enhancement, a hypersignal on diffusion-weighted imaging (DWI), and restricted signaling on the attenuated diffusion coefficient (ADC).[3]

Upon suspicion of a BA, a surgical procedure is warranted—either a puncture or excision—as soon as possible for both diagnostic and therapeutic purposes, according to recent European guidelines.[4] Despite high sensitivity and specificity of these MRI features highlighted by recent metanalysis (respectively 92 and 91%),[4] false negative represents ∼1 case out of 10[4] where abscess mimics other conditions such as metastasis. This carries a potential impact in timing of surgery and surgical planning. This is confirmed by Newman's recent systematic review on diagnostic errors in the emergency department,[5] where he found that intracranial abscesses are among the top 20 conditions associated with diagnostic errors in emergency departments and are ranked 11th in clinical conditions associated with serious misdiagnosis-related harms.

In this context, we conducted a systematic review including case reports and series, where the primary goal was to highlight situations where an abscess was mistaken for brain tumors. Plus, we provide an illustrative case.

Materials and Methods

Study Design

We conducted a systematic review according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guideline on MEDLINE/PubMed from inception to May 2024. We set out to systematically compile cases of BAs that were mistaken for brain tumors on imaging, and we added an illustrative case from our institution. Hence, we included all English-language articles with individual extractable data concerning cases of patients with BAs that were initially diagnosed as tumors on MRI scan. We excluded articles without English-written text or abstract, articles not directly relevant to the subject, and articles with significant lack of data.

Institutional review board (IRB) was not required for this study. A preliminary search based on PubMed, Cochrane Library, and Google Scholar revealed no previous systematic review on this specific subject. This study did not receive any funding or financial support. The authors have no personal, financial, or institutional interest related to this article.

Search Strategy

A search strategy was determined based on Cochrane handbook for systematic review.[6] First, we defined medical subject headings (MeSH) as follow: “brain”; “abscess”; “mimicking”; “tumor”; “metastasis”; “cancer”; “glioma”; “glioblastoma”; “case report”; “case series.” Then, we conducted a comprehensive literature search through MEDLINE/PubMed (https://pubmed.ncbi.nlm.nih.gov/) using the advanced search mode from inception to May 1, 2024. [Table 1] sums up MeSH and search terms used in MEDLINE/PubMed.

Table 1 Search terms used in electronic database

Database	MeSH and search terms
MEDLINE/Pubmed	MeSH: “brain”; “abscess”; “mimicking” ; “tumor” ; “metastasis”; “cancer”; “glioma”; “glioblastoma”; “case report”; “case series”
	Search Term: (brain[title]) AND ((abscess[title]) OR (abscesses[title])) AND (mimicking[title]) AND ((tumor[title]) OR (metastasis[title]) OR (glioma[title]) OR (glioblastoma[title]) OR (cancer[title]))

Data Collection

The data item list included general information related to the article (first author, year of publication, country of the first author and number of cases), the patients (gender, age, origin of the patient, immunodeficiency, and past medical history of cancer), clinical and radiological information (location of the lesion, MRI scan features, number of brain lesions, associated extracranial lesions), surgical management, medical treatment, type of germ, and outcome at 1 year.

Primary and Secondary Endpoints

The primary endpoint of the study was to highlight potential impact on the management of abscesses that were mistaken for tumors on brain MRI and determine the abscesses mimicking brain tumors. The secondary endpoints were to describe the scope of germs involved, the patients' characteristics, and the overall outcome at 1 year. No statistics were performed.

Results

Database Research

A systematic literature review was conducted to identify instances where BAs were misdiagnosed as brain tumors. Following the prescribed methods, the search across various databases yielded 20 records, with an additional case contributed by this study. After initial screening, two articles were excluded, and four more were discarded following a full-text review for eligibility. Ultimately, 14 studies were included[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] in the final review, plus our institutional case ([Fig. 1]), comprising a total of 15 cases ([Fig. 2]).

Fig. 1 Medline-based systematic literature review of patients presenting with brain abscess mistaken for brain tumors. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart. (Adapted from: Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71)

Fig. 2 Initial (A–D) and 1-year follow-up (E–H) contrast-enhanced brain MRI showing a right thalamus ring-enhancing lesion with restricted diffusion, suggestive of brain tumor: 59-year-old male, seropositive for HIV, after a 6-month antiviral therapy discontinuation, complaining of gait disorders, intermittent headaches, and a significant weight loss. Clinical examination revealed left-sided weakness (mMRC scale 4 out of 5 (4/5)). Laboratory tests indicated humoral immunodeficiency and a high viral load. Initial neuroimaging (A through D) revealed a ring-enhanced lesion in the right thalamus. Initially suggestive of a brain tumor, corticosteroid therapy was introduced. Stereotactic biopsy was performed, then empirical treatment with sulfadiazine and pyrimethamine was initiated, followed by ceftriaxone and metronidazole. Finally, polymerase chain reaction revealed an infection caused by Toxoplasma gondii. After a well-managed infectious treatment, at 1-year follow-up, the patient showed no motor or verbal neurological deficits. He was placed on trimethoprim-sulfamethoxazole. Brain MRIs at 1-year follow-up (E through H) show a residual contrast-enhanced lesion. (A) Initial contrast-enhanced axial T1-WI image; (B) initial contrast-enhanced sagittal T1-WI image; (C) initial diffusion-weighted image; (D) initial apparent diffusion coefficient; (E) 1-year follow-up contrast-enhanced axial T1-WI image; (F) 1-year follow-up contrast-enhanced sagittal T1-WI image; (G) 1-year follow-up diffusion-weighted image; (H) 1-year follow-up apparent diffusion coefficient. HIV, human immunodeficiency virus; mMRC, Modified Medical Research Council; MRI, magnetic resonance imaging.

Demographics and Clinical Presentation

The mean age of the patients was 59 years (ranging from 39 to 77 years). The male-to-female ratio was 2.8, with 11 males and 4 females ([Table 2]). Forty-seven percent of the patients (7 patients) had comorbidities related to induced immunosuppression, including HIV, glioblastoma, breast cancer, arthritis, gastric cancer, and nephrotic syndrome; all received adequate treatment. Initial symptoms included at least one focal deficit in 87% of patients (13/15), 27% of patients (4/15) reported generalized tonic-clonic seizures, signs of meningism were present in 13% of cases (2/15), likewise fever was present in 13% of patients (2/15), and 20% of cases (3/15) had confusion. One patient had a whole-body skin rash that resolved spontaneously ([Table 2]).

Table 2 Medline-based systematic literature review of patient presenting with brain abscess mistaken for brain tumors on imaging

Author, year, country	Age sex	Comorbidity	Treatment	Initial symptoms	MRI	Location	Extracranial-associated lesion	Surgery	Pathogen	Germ	Anti-biotherapy	GOS-E at 1 year
Choucha et al, 2024, France (present study)	♂ 59	HIV	Antiviral therapy in discontinuation for 6 months	Left side weakness Weight Loss	DWI: restricted diffusion ADC: hypersignal	Ring enhancing, right thalamus lesion	–	SB	Parasite	Toxoplasma gondii	Sulfadiazine Pyrimethamine Ceftriaxone Metronidazole	8
Kourouma et al, 2023, France[7]	♂ 77	GB	PrednisoneRadiation therapy Temozolomide	Left hemiparesis Dysexecutive syndrome	DWI: restricted diffusion ; PWI: no increased relative cerebral blood volume	Right polycystic frontal lobe (distant from GB)	–	B	Bacteria	Nocardia ignorata	ImipenemCotrimoxazole	8
Jung et al, 2023, Korea[8]	♀ 64	–	–	GTCS	DWI: restricted diffusion	Right temporal lobe	–	CR	Bacteria	Myco-bacterium tuberculosis	ClarithromycinAmikacinImipenem	8
Karan et al, 2019, Serbia[9]	♂ 70	–	–	Persistent fever and coughGTCS Right hemiplegia	DWI: restricted diffusion	Left precentral	Right middle lobe of the lung (6 cm)	CR	Bacteria	Nocardia cyriacigeorgica	SulfamethoxazoleCeftriaxone	7
De Laurentis et al, 2019, Italy[10]	♂ 54	–	–	HeadacheEmesisLeft hemianopsia	Central necrotic Ring enhancementImportant edema	Right medial temporo-occipital (30 × 28 × 34.5 mm)	–	CR (with 5-ALA)	Bacteria	Aggregatibacter aphrophilus	Ceftriaxone Metronidazole	6
Jang et al, 2018, Korea[11]	♂ 52	–	–	Whole body skin rash that lasted 1 month and resolved spontaneouslyPersisting headaches	T1 enhancement Central cyst	Peri-ventricular (20 mm)	–	SB	Bacteria	Rickettsia typhi	CeftazidimeVancomycin Ampicillin/Sulbactam	8
Khullar et al, 2015, India[12]	♀ 61	Breast cancer	Chemotherapy Hormonotherapy	Chronic headache FeverAltered behavior GTCSMeningismRight hemiparesis (3/5)	DWI: restricted diffusion	Left frontal lobulated lesion (30 × 29 mm)	–	CR	Bacteria	Gram-positive cocci	–	–
Asano et al, 2013, Japan[13]	♀ 76	–	–	StaggeringDysarthriaAppetite loss	DWI: hyperintensity	Disseminated multiple ring-enhanced brain masses	Left lower lobe of the lung	–	Bacteria	Nocardia	Sulfamethoxazole TrimethoprimCeftriaxone	8
Ouriemchi et al, 2011, France[14]	♂ 54	–	–	Right hemiparesis	CT: hypodense	Left capsulo-lenticular	–	SB	Bacteria	Arcanobacterium haemolyticum	Amoxicillin Metronidazole	7
Cianfoni et al, 2010, Italy[15]	♂ 67	–	–	Right hemiparesis	DWI: hyperintensity ADC: restriction PWI: increased rCBV T2WI: hyperintense Ring-enhanced lesion	Left parietal lobulated	–	CR	Bacteria	Nocardia	–	–
Damek et al, 2008, United States[16]	♀	GB	Radiation therapy Temozolomide Chemotherapy Prednisone	–	–	–	–	B	Fungus	Aspergillus terreus
Iannotti et al, 2008, United States[17]	♂ 53	–	–	Aphasia Confusion	Gadolinium-enhanced T1-weighted ring-enhancing	Left temporo-parietal multilobular (2.0 cm)	2.5 × 1.0 cm calcified nodule in the right lower lung lobe	CR	Bacteria	Nocardia farcinica	Trimethoprim Sulfamethoxazole	8
Yamada et al, 2005[18]	♀ 58	Arthritis	Prednisone	Aphasia Right hemiparesis	T1WI ring-enhanced lesion T2WI severe edema DWI: high intensity	Left frontal lobe	–	CR	Bacteria	Nocardia	Panipenem Betamipron Dexamethasone Phenytoin	8
Yokoyama et al, 2005, Japan[19]	♂ 57	Gastric cancer	–	GTCS Confusion	CT seems metastasis	Right parietal lobe	Lung: left superior lobe and bilateral inferior lobes	–	No germ found	–	–	1
Zhao et al, 2022, China[20]	♂ 39	Nephrotic syndrome	Cyclo-phosphamide Prednisone	Aphasia Confusion	T1W1: hypointensity in the left basal ganglia and the temporal lobe T2WI and FLAIR: hyperintense in the left midbrain, basal ganglia, and temporal-parietal lobe	Left basal ganglia Temporal lobe Midbrain	Upper lobe of the right lung	–	Bacteria	Providencia rettgeri	Ampicillin	7

Abbreviations: B, biopsy; CR, complete removal; CT, computed tomography; GB, glioblastoma IDH wild type; GOS-E, Glasgow Outcome Scale-Extended; GTCS, generalized tonic and clonic seizures; SB, stereotactic biopsy.

Imaging Features

Most patients (80% or 12 out of 15) initially underwent a brain MRI for assessment, while the remaining 3 patients (20%) had only a brain computed tomography (CT) scan (justified by the context of emergency in two cases, and one was published in the early 2000s, before the widespread adoption of MRI). In T1-weighted MRI sequences, two patients (13.3%) displayed more than one brain lesion. Among the other patients, lesions were categorized as follows: eight (61%) had a single ring-enhancing lesion, two (15%) had a multiloculated lesion, and three (20%) had multiloculated lesions. DWI frequently showed restricted diffusion in 46.7% of cases (7/15), while the ADC imaging indicated hyperintensity in 13.3% of cases (2/15). Perfusion-weighted imaging (PWI) was conducted in 13.3% of cases (2/15), revealing no increased cerebral blood volume in either. T2-weighted images and T2 fluid-attenuated inversion recovery sequences were commonly hyperintense.

Lesions were located in diverse brain regions, including the thalamus, frontal lobe, temporal lobe, precentral area, temporo-occipital region, periventricular area, capsulo-lenticular area, parietal lobe, basal ganglia, midbrain, and temporoparietal lobe. Additionally, extracranial lesions were identified in five cases, notably in the lungs.

Surgical Management and Pathogens

In five cases (33%), a biopsy or stereotactic biopsy was performed. In seven cases (47%), open surgery was conducted with complete removal of the lesion, via extracapsular excision of BA; the surgical modality was not reported for three cases ([Table 2]). Pathogens were identified in 14 cases (93%). Most of these pathogens (12 cases, 80%) were bacterial. Specifically, 40% of cases (6/15) were attributable to Nocardia, and one case each was caused by Mycobacterium, Aggregatibacter, Rickettsia, Arcanobacterium, Providencia, and one unidentified gram-negative cocci ([Table 3]). Additionally, one infection was due to the fungus Aspergillus terreus, and another was caused by the parasite Toxoplasma gondii.

Table 3 Pathogen involved

Pathogen	Number of cases (percentage)
Bacteria	12 (80%)
Nocardia Ignorata Cyriacigeorgica Farcinica Spp.	6 (40%) 1 1 1 3
Mycobacterium tuberculosis	1
Aggregatibacter aphrophilus	1
Rickettsia typhi	1
Arcanobacterium haemolyticum	1
Providencia rettgeri	1
Cocci Gram (no details)	1
Parasite Toxoplasma gondii	1 (7%) 1
Fungus Aspergillus terreus	1 (7%) 1
Not found	1 (7%)

Treatments and Outcomes

Treatments and outcomes varied significantly, influenced by the type of comorbidities and pathogens involved. Antiviral therapy was used in one case, while antibiotics were more commonly prescribed, including sulfadiazine, pyrimethamine, ceftriaxone, metronidazole, imipenem, clarithromycin, amikacin, co-trimoxazole, ceftazidime, vancomycin, ampicillin/sulbactam, trimethoprim, and sulfonamides.

Outcomes were assessed using the Glasgow Outcome Scale-Extended (GOS-E) at 1 year, with scores ranging from 1 to 8. The average GOS-E score was ∼7, indicating generally good recovery. Specifically, seven patients achieved a top score of 8, indicating excellent recovery. Additionally, two patients scored 7, and one each scored 6, 4, and 2. The lowest recorded score was 1. One patient, followed for only 1 month, did not have an applicable GOS-E score.

Discussion

Imaging

Imaging plays a central role in the management of BAs.[3] Latest European guidelines strongly recommend contrast-enhanced brain MRI with DWI and ADC sequences for the screening of BA.[4] When MRI is unavailable, contrast-enhanced brain CT is recommended as an alternative.[4] Imaging features suggestive of BAs typically include ring-enhancement, hypersignal in DWI, and restriction in ADC, with peripheral edema.[3]

According to a recent meta-analysis,[4] these features can help distinguishing BAs from neoplastic lesions with high sensitivity and specificity (respectively 92 and 91%), with a positive and negative predictive value of respectively 88 and 90%.[4] [21] Conversely, this means that approximately 1 out of 10 patients is a false negative and may be misidentified as having a brain tumor. This can be due to the composition of the abscess which can evolve over time or depend on factors such as pathogen, immune status, or antibiotic therapy.

This misdiagnosis can potentially affect the timing of surgery, surgical planning, and the time to antibiotic therapy. Although, to the best of our knowledge, no study directly correlates the timing of surgery with patient outcomes, it is reasonable to hypothesize that prolonged delays in surgical and antibiotic interventions may increase the risks of abscess rupture, higher case-fatality rates, or neurological sequelae. This concern is highlighted by a systematic review on diagnostic errors in emergency departments, which identifies missed intracranial abscesses as the 11th leading cause of serious misdiagnosis-related harms.[5]

Potential Causes of False Negative

Based on our review and current literature, several factors may explain the 10% rate of false negatives, although we cannot conclusively affirm these due to limited statistical power and the lack of a comparative group.

First, we observed a high prevalence of immunologic deficits; 47% of reported cases had either a direct cause, such as HIV, or an indirect cause through treatments like steroids or chemotherapy, which could potentially affect the quality of pus and influence DWI diffusion sequence imaging.

Second, the type of causative pathogen appears to impact imaging features. Our review found that most pathogens were bacteria typically transmitted through hematogenous dissemination, often seen in immunocompromised patients. Notably, Nocardia species stood out, accounting for 40% of cases. A recent systematic review of cases of BA secondary to Nocardia noted that half of the 54 BA cases displayed an atypical hyposignal in DWI at the center of lesions.[22] Further studies are needed to explore why this bacterium is associated with such a significant number of reported cases and whether these imaging features are directly linked to the germ or the immunocompromised status of the patients.

Other risk factors identified include prior antibiotic therapy for several weeks and post-neurosurgical BAs.[4] [23] These aspects require more robust analysis with greater statistical power to determine their definitive impact on imaging outcomes.

Perspective: Toward New MRI Protocols

In addition to the traditional MRI protocol implemented two decades ago (described above), research in the field of radiology is performed to address its limitations.

The intralesional susceptibility signal (ILSS) in susceptibility-weighted imaging (SWI) is a potential MRI sequence that could contribute to the diagnosis of BA. According to Fu et al's series, a combined protocol with ADC and ILSS offers a strong diagnostic accuracy to differentiate glioblastoma from abscess; however, this protocol performed equally in differentiating BA and metastasis.[24] [25] SWI has been confirmed as one of the most sensitive MRI sequences for the detection of blood products and calcifications within brain lesions. In fact, a very recent study testifies how SWI, which can be acquired both with and without gadolinium contrast, has proven to be particularly useful in the diagnosis, characterization, and grading of gliomas, as it reveals intratumoral susceptibility signals (ILSS) indicative of microhemorrhages, calcifications, or other paramagnetic elements. The presence and pattern of ILSS in SWI can provide essential diagnostic clues that help distinguish gliomas from other types of brain lesions, including abscesses and infections. However, challenges remain in refining the role of SWI in glioma evaluation, particularly in standardizing its interpretation and integrating it with other imaging modalities. This highlights the potential of SWI not only for glioma management but also as a supportive tool in complex cases where traditional imaging modalities provide limited information.[26]

First, regarding in vivo 1H-MR spectroscopy, in a retrospective study of 194 cases amino acid resonance (AAs) with or without other metabolites was observed in 80% of abscesses, with a sensitivity and specificity of 0.72 and 0.30, respectively. Most obligate anaerobes and some facultative anaerobes showed the presence of Lac/Lip, Aas, and Acs with or without Suc. Most obligate aerobes or facultative anaerobes showed the presence of Lac and AAs, with or without lipids.[27]

Second, perfusion sequences, especially relative cerebral blood flow (rCBV), may be useful.[28] Indeed, BA's capsule is poorly vascularized on PWI-MRI, compared with the tumor's wall of metastasis, which is more likely to display hypervascularization.[29] Notably, MR spectroscopy is neither sensitive nor specific to differentiate cerebral abscesses from brain metastases.[22]

Additionally, new computational tools such as machine learning are emerging, and have been leveraged to enhance diagnostic accuracy. Notably, an efficient artificial intelligence model, based on radiomics techniques and developed through a cohort of 186 patients, demonstrated improved predictive capabilities in distinguishing cystic brain metastases from abscesses.[30]

Based on the current literature, we do not advocate for the widespread adoption of any of these sequences or tools now. Indeed, these sequences are currently still evaluated, and their indications should be balanced with the expertise of local clinicians, logistics, and the patients' specific features. Yet, none seems to offer overwhelming sensitivity or specificity over the usual protocol.

Important Lessons and Key Messages

Clinicians must bear in mind that the current MRI protocol, despite its 90% sensitivity and specificity in diagnosing BAs, still carries a 10% false-negative rate. Hence, any patient presenting with an expansive process, clinical indications of a potential abscess, and restricted diffusion on MRI should be considered as suffering either from a brain tumor or a BA and should be managed as a neurosurgical emergency accordingly.

Clinical clues guiding toward BA include meningism, fever, or clinical deterioration not related to a sudden increase in brain edema. Special attention should be given to patients with known or suspected immune deficiencies, those recently treated with antibiotics for other infections, or those with associated infections. Notably the classic triad of symptoms present in patients having BA related to Nocardia, namely headache, fever, and focal deficit, has been reported in less than 30% of cases.[22]

Initial diagnostic steps should include conducting blood tests to investigate further signs of an infection. Notably, HIV-positive patients should have toxoplasmosis being ruled out. Additionally, for patients suspected of having a brain tumor that might be a false negative for an abscess on the current MRI protocol, it is advisable to consider additional MRI sequences, depending on the center's facilities and local competence, such as the ILSS-SWI or rCBV measurement.

A large-scale Danish study found that delaying antibiotics until aspiration or excision did not increase the 6-month mortality risk. Besides, empirical antibiotic treatment can lead to false-negative microbiological cultures,[31] especially for very fragile surgical samples such as BA.[30] Hence such a scenario requires emergency surgery both for diagnostic and therapeutic purposes. European guidelines recommend that patients suspected of having a BA should undergo surgery within 24 hours of the MRI. Samples from the lesion should be sent to both pathology and bacteriology departments and the bacteriology department should be notified to check for rare bacteria like Nocardia. Post-surgical empirical antibiotics should be considered on a case-by-case basis.

Limitations

Our study faces inherent bias due to its design. First, a publication bias as vast amount of clinical cases may have not been published regardless of the outcome. Second, we acknowledge the presence of a selection bias in our study due to the specific keywords used. We focused on including cases of BAs that mimicked cerebral tumors, requiring this information to be highlighted in the title as one of the key messages. Consequently, numerous cases of BAs resembling cerebral tumors that were reported in various other series may have been inadvertently omitted. Third, some information which could have been useful were missing, such as the delay between imaging and surgery. Indeed, it would have helped measuring the impact of this imaging feature on the management. Finally, we do not have enough data to conduct conclusive statistical analysis.

Conclusion

In conclusion, BA can be challenging to diagnose through MRI with a notable 10% rate of false negative. This seems especially to be the case for immunocompromised patients, patients who had several weeks of antibiotic therapy prior to imaging, or secondary to certain pathogen such as nocardia or mycobacterium tuberculosis. Improved MRI protocols with techniques such as SWI and machine learning have enhanced our ability to differentiate abscesses from neoplastic lesions. Future research should focus on refining diagnostic criteria and incorporating a broader range of imaging features.

Conflict of Interest

None declared.

Authors' Contributions

The study was conceptualized by A.C. and M.D.S., with methodology developed by N.B. and H.D. Formal analysis and investigation were conducted by A.C. and M.D.S., while the original draft was prepared collaboratively by A.C., M.D.S., N.B., and S.H. J.-C.L. and H.D. contributed to the review and editing of the manuscript. Resources were provided by A.C. and M.D.S., and supervision was carried out by J.-C.L. and H.D.

Ethical Approval

All principles in the Declaration of Helsinki were followed, and the Italian laws on privacy (Art. 20–21, DL 196/2003) as published in the Official Journal, volume 190, August 14, 2004, which explicitly waives the need for ethical approval for the use of anonymous data, were respected.

• References

• 1 Bodilsen J, Dalager-Pedersen M, van de Beek D, Brouwer MC, Nielsen H. Incidence and mortality of brain abscess in Denmark: a nationwide population-based study. Clin Microbiol Infect 2020; 26 (01) 95-100
• 2 Bodilsen J, Dalager-Pedersen M, van de Beek D, Brouwer MC, Nielsen H. Long-term mortality and epilepsy in patients after brain abscess: a nationwide population-based matched cohort study. Clin Infect Dis 2020; 71 (11) 2825-2832
• 3 Reddy JS, Mishra AM, Behari S. et al. The role of diffusion-weighted imaging in the differential diagnosis of intracranial cystic mass lesions: a report of 147 lesions. Surg Neurol 2006; 66 (03) 246-250 , discussion 250–251
• 4 Bodilsen J, D'Alessandris QG, Humphreys H. et al; ESCMID Study Group for Infections of the Brain (ESGIB). European Society of Clinical Microbiology and Infectious Diseases guidelines on diagnosis and treatment of brain abscess in children and adults. Clin Microbiol Infect 2024; 30 (01) 66-89
• 5 Newman-Toker DE, Peterson SM, Badihian S. et al. Diagnostic Errors in the Emergency Department: A Systematic Review [Internet]. Rockville, MD: Agency for Healthcare Research and Quality (AHRQ); 2022
• 6 Higgins JP, Green S. eds. Cochrane Handbook for Systematic Reviews of Interventions: Cochrane Book Series. 1st ed. Chichester, UK: Wiley; 2008
• 7 Kourouma M, Joubert B, Berhouma M, Roux S, Meyronet D, Ducray F. Nocardia brain abscess mimicking tumor progression in an elderly glioblastoma patient treated with temozolomide radiochemotherapy. Rev Neurol (Paris) 2023; 179 (08) 929-931
• 8 Jung J, Shin I, Choi Y. A rare case of nontuberculous mycobacterial abscess mimicking brain tumor in an immunocompetent patient. Brain Tumor Res Treat 2023; 11 (03) 219-222
• 9 Karan M, Vučković N, Vuleković P, Rotim A, Lasica N, Rasulić L. Nocardial brain abscess mimicking lung cancer metastasis in immunocompetent patient with pulmonary nocardiasis: a case report. Acta Clin Croat 2019; 58 (03) 540-545
• 10 de Laurentis C, Del Bene M, Fociani P, Tonello C, Pollo B, DiMeco F. 5-ALA fluorescence in case of brain abscess by aggregatibacter mimicking glioblastoma. World Neurosurg 2019; 125: 175-178
• 11 Jang Y, Moon J, Jun J-S. et al. Case of Rickettsia Typhi-induced brain abscess mimicking brain tumor. Osong Public Health Res Perspect 2018; 9: 122-125
• 12 Khullar P, Datta NR, Wahi IK, Kataria S. Brain abscess mimicking brain metastasis in breast cancer. J Egypt Natl Canc Inst 2016; 28 (01) 59-61
• 13 Asano M, Fujimoto N, Fuchimoto Y. et al. Brain abscess mimicking lung cancer metastases; a case report. Clin Imaging 2013; 37 (01) 147-150
• 14 Ouriemchi W, Jeddi D, Ziane Y, El Quessar A, Benouda A. Abcès cérébral àArcanobacterium haemolyticum mimant une tumeur cérébrale. Med Mal Infect 2011; 41 (07) 397-399
• 15 Cianfoni A, Calandrelli R, De Bonis P, Pompucci A, Lauriola L, Colosimo C. Nocardia brain abscess mimicking high-grade necrotic tumor on perfusion MRI. J Clin Neurosci 2010; 17 (08) 1080-1082
• 16 Damek DM, Lillehei KO, Kleinschmidt-DeMasters BK. Aspergillus terreus brain abscess mimicking tumor progression in a patient with treated glioblastoma multiforme. Clin Neuropathol 2008; 27 (06) 400-407
• 17 Iannotti CA, Hall GS, Procop GW, Tuohy MJ, Staugaitis SM, Weil RJ. Solitary Nocardia farcinica brain abscess in an immunocompetent adult mimicking metastatic brain tumor: rapid diagnosis by pyrosequencing and successful treatment. Surg Neurol 2009; 72 (01) 74-79 , discussion 79
• 18 Yamada SM, Nakai E, Toyonaga S, Nakabayashi H, Park KC, Shimizu K. A rapidly enlarging nocardial brain abscess mimicking malignant glioma. J Nippon Med Sch 2005; 72 (05) 308-311
• 19 Yokoyama T, Hyodo M, Hosoya Y. et al. Aggressive G-CSF-producing gastric cancer complicated by lung and brain abscesses, mimicking metastases. Gastric Cancer 2005; 8 (03) 198-201
• 20 Zhao Y, Lian B, Liu X. et al. Case report: cryptogenic giant brain abscess caused by Providencia rettgeri mimicking stroke and tumor in a patient with impaired immunity. Front Neurol 2022; 13: 1007435
• 21 Siddiqui H, Vakil S, Hassan M. Diagnostic accuracy of echo-planar diffusion-weighted imaging in the diagnosis of intra-cerebral abscess by taking histopathological findings as the gold standard. Cureus 2019; 11 (05) e4677
• 22 Beucler N, Farah K, Choucha A. et al. Nocardia farcinica cerebral abscess: a systematic review of treatment strategies. Neurochirurgie 2022; 68 (01) 94-101
• 23 Farrell CJ, Hoh BL, Pisculli ML, Henson JW, Barker II FG, Curry Jr WT. Limitations of diffusion-weighted imaging in the diagnosis of postoperative infections. Neurosurgery 2008; 62 (03) 577-583 , discussion 577–583
• 24 Fu J-H, Chuang T-C, Chung H-W. et al. Discriminating pyogenic brain abscesses, necrotic glioblastomas, and necrotic metastatic brain tumors by means of susceptibility-weighted imaging. Eur Radiol 2015; 25 (05) 1413-1420
• 25 Lai P-H, Chung H-W, Chang H-C. et al. Susceptibility-weighted imaging provides complementary value to diffusion-weighted imaging in the differentiation between pyogenic brain abscesses, necrotic glioblastomas, and necrotic metastatic brain tumors. Eur J Radiol 2019; 117: 56-61
• 26 Martín-Noguerol T, Santos-Armentia E, Ramos A, Luna A. An update on susceptibility-weighted imaging in brain gliomas. Eur Radiol 2024; 34 (10) 6763-6775
• 27 Pal D, Bhattacharyya A, Husain M, Prasad KN, Pandey CM, Gupta RK. In vivo proton MR spectroscopy evaluation of pyogenic brain abscesses: a report of 194 cases. AJNR Am J Neuroradiol 2010; 31 (02) 360-366
• 28 De Simone M, Fontanella MM, Choucha A. et al. Current and future applications of arterial spin labeling MRI in cerebral arteriovenous malformations. Biomedicines 2024; 12 (04) 753
• 29 Erdogan C, Hakyemez B, Yildirim N, Parlak M. Brain abscess and cystic brain tumor: discrimination with dynamic susceptibility contrast perfusion-weighted MRI. J Comput Assist Tomogr 2005; 29 (05) 663-667
• 30 Cui L, Qin Z, Sun S, Feng W, Hou M, Yu D. Diffusion-weighted imaging-based radiomics model using automatic machine learning to differentiate cerebral cystic metastases from brain abscesses. J Cancer Res Clin Oncol 2024; 150 (03) 132
• 31 Smith SJ, Ughratdar I, MacArthur DC. Never go to sleep on undrained pus: a retrospective review of surgery for intraparenchymal cerebral abscess. Br J Neurosurg 2009; 23 (04) 412-417

Publication History

Article published online:
06 February 2025