Proposed computed tomography severity index for the evaluation of invasive fungal sinusitis: Preliminary results

Abstract

BACKGROUND

Invasive fungal sinusitis (IFS) can present as a mild disease to life-threatening infection. A recent surge in cases was seen due to the coronavirus disease 2019 (COVID-19) pandemic. Many patients require surgical debridement and hence imaging [contrast-enhanced computed tomography (CECT) of the paranasal sinuses (PNS)] to document the extent of the disease. However, there was no scoring system using CECT to describe the severity of IFS. This study proposes a computed tomography (CT) severity index (CTSI) to describe the severity of rhino-orbital-cerebral involvement in symptomatic COVID-19 patients and hypothesizes that higher CTSI correlates with disease severity and thus slow response/non-response to treatment.

AIM

To propose a scoring system using CECT to describe the severity of IFS and correlate it with clinical outcomes.

METHODS

A prospective study on 66 COVID-19 positive patients with CECT PNS done for IFS was performed. Split-bolus single-phase CT technique was used. Based on the extent of involvement, a CTSI was designed. Disease in four major subsite areas was assessed. Each subsite involvement was given points according to this model and then summated. Based on the final summated CTSI, the disease was classified as mild, moderate, or severe. Two subsets were subsequently analyzed including survival and death; and responders and non-responders.

RESULTS

The study cohort was 66 COVID-19-positive patients with suspected IFS with a median age of 48.5 years. Mild disease was noted in 34 (51.52%), moderate in 28 (42.42%), and severe disease in 4 (6.06%) patients. There was a significant association of mortality and poor clinical response (P = 0.02) with disease bilaterality. Laterality and CTSI were significant predictors of response to treatment. The mean CTSI of responders was 6.3, of non-responders was 12.9 and the response to treatment was significantly associated with CTSI (t-test, P < 0.001). Receiver operating characteristic curve analysis (Liu method) to distinguish between responders and non-responders showed that the cut-off value for CTSI of 11 had a sensitivity of 78.26% and a specificity of 95.35% to predict response assessment.

CONCLUSION

CTSI can help in quantification of the disease burden, mapping out disease extent, triaging patients, and response assessment; especially patients with underlying comorbidities. A higher score would alert the clinician to initiate aggressive treatment, as severe disease correlates with slow response/non-response to the treatment.

Core Tip: This was a prospective study with 66 coronavirus disease 2019-positive patients in whom contrast-enhanced computed tomography (CECT) of the paranasal sinuses was done for invasive fungal sinusitis (IFS). It aimed to propose a scoring system [computed tomography severity index (CTSI)] using CECT to describe the severity of IFS and correlate it with clinical outcomes. Disease was categorized as mild, moderate, and severe based on CTSI. A higher score correlated with slow response/non-response to treatment in our study. Thus, CTSI can help in the quantification of the disease burden, triaging patients, and response assessment. A higher score would alert the clinician to initiate aggressive treatment, as severe disease correlates with slow response/non-response to treatment.

INTRODUCTION

Invasive fungal sinusitis (IFS) due to mucormycosis/aspergillus can present with a clinical spectrum ranging from mild disease to life-threatening infection. Clinically, IFS can present with atypical signs and symptoms like complicated sinusitis, such as nasal blockage, crusting, proptosis, facial pain and edema, ptosis, chemosis, and even ophthalmoplegia[1]. Headache, fever, and various neurological signs and symptoms are seen if intracranial extension is present. Without early diagnosis and treatment, there may be a rapid progression of the disease, with reported mortality rates from intra-orbital and intracranial complications of 50-80 percent[2]. Many patients require surgical debridement and hence imaging to document the extent of the disease[3]. Contrast-enhanced computed tomography (CECT) of the paranasal sinuses (PNS) is the ideal modality in the emergency setting for adequate demonstration of the extra-sinus spread of disease to orbit, infratemporal fossa, skull base, and intracranial involvement along with bony erosions.

In the recent past, there was a pandemic due to coronavirus disease 2019 (COVID-19) and an associated rise in the cases of IFS. This was because COVID-19 infection and its treatment (mostly with steroids) predispose patients to many opportunistic infections at multiple sites including paranasal sinuses[4,5].

There has been no scoring system using CECT to describe the severity of IFS and its relation to the clinical severity in patients with COVID-19.

Therefore, we undertook this study to propose a computed tomography (CT) severity index (CTSI) to describe the severity of rhino-orbital-cerebral involvement in symptomatic COVID-19 patients using CECT paranasal sinuses and to correlate the radiological changes with the clinical outcomes.

MATERIALS AND METHODS

Ninety-five patients with diagnosed COVID-19 and suspected IFS underwent CT PNS during the study period of April 2021 to January 2022. Non-contrast scans were acquired for fifteen patients with deranged kidney function tests and were excluded from the study group. Fourteen other patients were also excluded due to inadequate information or follow-up. Thus, data from 66 COVID-19-positive patients with CECT PNS done for IFS was analysed (Figure 1).

Figure 1.

Flowchart to depict the methodology of the study. IFS: Invasive fungal sinusitis; CT: Computed tomography; PNS: Paranasal sinuses; CECT: Contrast-enhanced computed tomography; NCCT: Non-contrast computed tomography.

CT acquisition

Split-bolus single-phase CT technique[6] was used to optimally demonstrate the vascular (arterial and venous) as well as mucosal extent of the disease. This protocol was performed with 1 mL/kg of non-ionic contrast medium of concentration 300 mg iodine/mL (Iohexol) administered through an 18 to 20 G cannula in the antecubital vein using a pressure injector. Two-thirds of the contrast was given, 4 mL/sec followed by 30 seconds gap and then the rest of the contrast was given, 3.5 mL/sec followed by automatic bolus tracking and image acquisition. With the use of this technique, there was simultaneous opacification of the internal carotid artery and cavernous sinus along with optimal soft tissue demonstration of the extra-sinus spread of disease to orbit, infratemporal fossa, skull base, and intracranial involvement.

Image analysis

The scans were viewed on the reporting stations using multiplanar reconstructions in both soft tissue and bone windows. Further, a structured reporting layout was employed in each case to reflect the extent of the disease in a standardized format.

CTSI

Based on the extent of involvement, a CTSI was designed (Table 1). Four major areas included were disease in nose/paranasal sinuses; adjacent soft tissue infiltration; orbit and intracranial involvement. The involvement of each subsite was given points according to this model and then summated. For each patient, the final CTSI was calculated, and the disease was further classified as mild (CTSI 1-8), moderate (9-16) or severe (17-25). Demographic analysis was done, including clinical presentation and presence of co-morbidities like diabetes and hypertension. Two subsets were subsequently analyzed including survival and death; and responders and non-responders. Residual clinical disease/ worsening (on monthly nasal cavity cleaning and clinical assessment) or persistent disease in the follow-up scans (performed at 1-3 months intervals) were taken as criteria of non-responder. Clinical staging was done as per the TALMI staging system[7].

Table 1.

Computed tomography severity scoring index

Items	Value
Disease limited to nose/PNS
Mucosal disease in paranasal sinuses	1
Mucosal disease in nasal cavity and nasopharynx	1
Adjacent tissue invasion
Hard palate erosion	1
Soft tissue infiltration anterior/posterior periantral fat	1
Soft tissue infiltration extending to PPF/SOF/IOF/Orbital apex/ITF	2
Significant bilateral disease	2
Orbit
Soft tissue/fat/muscle/NLD involvement	1
Bone erosion	2
Intraocular/optic nerve involvement	3
Intracranial disease
Skull base invasion (erosion)	2
Cavernous sinus involvement	3
Internal carotid artery narrowing/occlusion	3
Intracranial complications (meningitis/cerebritis/abscess/infarct)	3

Mild: 1-8; Moderate: 9-16; Severe: 17-25. PNS: Paranasal sinuses; PPF: Pterygopalatine fossa; SOF: Superior orbital fissure; IOF: Inferior orbital fissure; ITF: Infratemporal fossa; NLD: Nasolacrimal duct.

Statistical analysis

Data was recorded on a pre-designed proforma and managed on an Excel spreadsheet. Statistical analysis was done using the Stata 14.1. A P value of less than 0.05 was considered significant.

RESULTS

Demographics

The study cohort was of 66 COVID-19-positive patients with suspected IFS. The age of the patients included in the study ranged from 26 to 71 years (median 48.5 years, Interquartile range 40-60.75 years). Males predominated the study cohort (male:female = 5.6:1) with most of the patients (69.70%) from the urban background. The clinical presentation included periorbital pain (75.76%), facial numbness (63.64%), periorbital swelling (56.06%), restricted extra-ocular movements (56.06%), reduced visual acuity (60.61%), proptosis (33.33%), palatal involvement (31.82%), ptosis (27.27%), and cranial nerve involvement (27.27%; Table 2). About three-fourths of the patients (n = 51) were known diabetic while a total of 56 patients had their glycosylated hemoglobin above 6.5 gm/dL (mean HbA1c of 9.11 ± 2.20 gm/dL). Thirty-five percent of patients were hypertensive. Six (9.09%) patients had bilateral disease.

Table 2.

Demographic details

	Parameter	Criteria	n (%)
1	Gender	Male:Female	56 (84.85):10 (15.15)
2	Periorbital swelling	Absent:Present	29 (43.94):37 (56.06)
3	Periorbital pain	Absent:Present	16 (24.24):50 (75.76)
4	Facial numbness	Absent:Present	24 (36.36):42 (63.64)
5	Extra-ocular movements	Normal:Restricted:Frozen	29 (43.94):20 (30.3):17 (25.76)
6	Vision status	Normal:Reduced:FCCF:PL Negative	26 (39.39):21 (31.82):1 (1.52):18 (27.27)
7	Proptosis	Absent:Present	44 (66.67):22 (33.33)
8	Palatal involvement	Absent:Present	45 (68.18):21 (31.82)
9	Cranial nerve Involvement	Absent:Present	48 (72.73):18 (27.27)
10	Ptosis	Absent:Present	39 (59.09):27 (40.91)
11	H/o diabetes	Absent:Present	15 (22.73):51 (77.27)
12	H/o hypertension	Absent:Present	43 (65.15):23 (34.85)
13	HbA1c	< 6.5:≥ 6.5	10 (15.15) 56 (84.85)
14	Side involved	Unilateral:Bilateral	60 (90.91):6 (9.09)
15	CT severity index	Mild (1-8):Moderate (9-16):Severe (≥ 17)	34 (51.52):28 (42.42):4 (6.06)
16	Final organism	Mucormycosis:Aspergillosis:Both	45 (68.18):9 (13.64):12 (18.18)
17	Clinical outcome	Responder:Non-responder	43 (65.15):23 (34.85)
18	Survival	Survived:Expired	56 (84.85):10 (15.15)
19	Diagnosis of COVID-19 before IFS	After:Before	9 (13.63):57 (86.36)
20	Environment	Rural:Urban	20 (30.3):46 (69.7)
21	TALMI staging	Stage 1:Stage 2:Stage 3:Stage 4	2 (3.03):31 (46.97):7 (10.61):26 (39.39)
22	Steroid given for COVID-19 (n = 57)	Not given:Given	24 (42.1):33 (57.89)

CT: Computed tomography; COVID-19: Coronavirus disease 2019.

The CTSI was calculated and grouped as mild (score 1-8), moderate (score 9-16), or severe (score ≥ 17). Mild disease was noted in 34 (51.52%), moderate in 28 (42.42%), and severe disease in 4 (6.06%) patients (Table 2). As per the TALMI clinical scoring systems, more patients had Stage 2 disease (46.97%).

Out of the 66 patients, 57 had COVID-19 before diagnosis of IFS, ranging from a period of 1-36 days. Nine patients were diagnosed to be COVID-19 positive during the admission for IFS. No statistically significant difference between the CTSI and TALMI staging was seen in these two subgroups using the Pearson χ² test.

Forty-five (68.18%) patients had fungi of the Mucorales order identified in the tissue specimen while Aspergillus sp. was identified in 9 (13.64%) patients as the sole causative agent. Twelve (18.18%) patients had coexistent infection. The difference in CTSI, survival rates and response to treatment between the subgroups of different etiological agents is summarised in Table 3. There was no statistically significant difference in the survival rates (P value = 1) or response to treatment (P value = 0.53) between the three subgroups.

Table 3.

Comparison of severity of disease in different etiological agents, n (%)

Parameter	Subclassification of parameter	Aspergillosis (n = 9)	Mucormycosis (n = 45)	Mixed (n = 12)
Diabetes mellitus		3 (33.3)	37 (82.2)	11 (91.7)
Hypertension		2 (22.2)	16 (35.6)	5 (41.7)
Laterality	Unilateral	9 (100)	40 (89.9)	11 (91.7)
	Bilateral	0 (0)	5 (10.1)	1 (8.3)
CTSI	Mild	7 (77.8)	21 (46.7)	6 (50)
	Moderate	2 (22.2)	21 (46.7)	5 (41.7)
	Severe	0 (0)	3 (6.7)	1 (8.3)
Survival	Survival	8 (88.9)	38 (84.4)	10 (83.3)
	Mortality	1 (11.1)	7 (15.6)	2 (16.7)
Responder/non-responder	Responder	7 (77.8)	27 (60)	9 (75)
	Non-responder	2 (22.2)	18 (40)	3 (25)

CTSI: Computed tomography severity index.

Correlation with mortality

There were ten deaths with a survival rate of 84.85%. Comparative analysis was done amongst survivors and patients who expired. There was no association found between survival and history of diabetes or hypertension, causative organism, CT severity score, or TALMI staging. However, there was a significant association of bilateral disease with mortality (Table 4). The level of glycosylated hemoglobin also had no significant effect on mortality (t-test, P = 0.79). Steroids were given to manage COVID-19 in 33 (57.89%) patients before onset of IFS. However, no significant correlation was observed between survival and steroid use.

Table 4.

Correlation with survival and death

Parameter	Criteria	Survivor	Expired	Total	P value (Fischer's exact test)
Diabetes mellitus	Absent	14	1	15
	Present	42	9	51	0.28
Hypertension	Absent	36	7	43
	Present	20	3	23	0.52
Laterality	Unilateral	54	6	60
	Bilateral	2	4	6	0.004
Organism	Mucormycosis	38	7	45
	Aspergillosis	8	1	9
	Mixed	10	2	12	1
Clinical response	Responder	38	5	43
	Non-responder	18	5	23	0.23
TALMI staging	Stage 1	2	0	2
	Stage 2	29	2	31
	Stage 3	5	2	7
	Stage 4	20	6	26	0.16
CTSI	Mild (1 to 8)	29	5	34
	Moderate (9 to 16)	25	3	28
	Severe (17 or more)	2	2	4	0.12

CTSI: Computed tomography severity index.

Responders vs non-responders

Clinico-radiological assessment was done for the follow-up of these patients. This was in the form of monthly nasal cavity cleaning and evaluation followed by radiology based on symptoms and signs. Residual clinical disease/ worsening and persistent disease in the follow-up scans (performed at 1-3 months intervals) were taken as criteria of non-responder. Forty-three patients showed clinico-radiological resolution whereas 23 patients were non-responders. Laterality and CTSI were significant predictors of response to treatment (Table 4). Out of the six patients who had bilateral disease, five were non-responders and bilaterality of disease had a significant association with poor clinical response (P = 0.02; Table 5). The mean CTSI of responders was 6.3, of non-responders was 12.9 and the response to treatment was significantly associated with CTSI (t-test, P < 0.001). On further categorizing the disease as mild, moderate, and severe on CTSI, it was seen that the majority (91%) of patients with mild disease were responders (Figure 2). Sixteen out of 28 patients (57%) with moderate disease and all the patients with severe disease (n = 4) were non-responders (Figure 3 and Figure 4). Receiver operating characteristic curve analysis (Figure 5; Liu method) to distinguish between responders and non-responders showed that the cut-off value for CTSI of 11 had a sensitivity of 78.26% and a specificity of 95.35% to predict response assessment.

Table 5.

Correlation with Responder and Non-responder.

Parameter	Criteria	Responder	Non-responder	Total	P value (Fischer's exact test)
Diabetes mellitus	Absent	12	3	15
	Present	31	20	51	0.14
Hypertension	Absent	27	16	43
	Present	16	7	23	0.39
Laterality	Unilateral	42	18	60
	Bilateral	1	5	6	0.02
Organism	Mucormycosis	27	18	45
	Aspergillosis	7	2	9
	Mixed	9	3	12	0.53
Survival	Survivor	38	18	56
	Expired	5	5	10	0.23
TALMI Staging	Stage 1	2	0	2
	Stage 2	22	9	31
	Stage 3	4	3	7
	Stage 4	15	11	26	0.58
CTSI	Mild (1 to 8)	31	3	34
	Moderate (9 to 16)	12	16	28
	Severe (17 or more)	0	4	4	0.00

CTSI: Computed tomography severity index.

Figure 2.

Mild disease with computed tomography severity index 2 (67-year-old male, coronavirus disease 2019 positive with mild headache and right-sided facial pain). The patient was a known diabetic with an HbA1C of 8.5. Microbiological evaluation revealed both Aspergillus sp. and Rhizopus sp. A-C: Axial computed tomography (CT) images show minimal soft tissue within the left ethmoid sinus (A) and right maxillary sinus infiltrating the adjacent right posterior retroantral fat (B). The bilateral eye globes and orbits are normal. There is e/o erosion in the anterior maxillary wall with soft tissue infiltrating the preantral space (C); D-F: Coronal bone window CT images. Figure D shows bone erosion in the anterior maxilla (arrow). There is evidence of bilateral disease with mild mucosal disease in the right maxillary sinus (dotted arrow in F) and left ethmoid sinus (dotted arrow in E) with mucosal thickening of the right inferior turbinate (arrow in F). The hard palate is intact (arrow in E). The total summated CT severity index for this case is 2 - mild disease (1 point for involvement of right maxillary sinus and ethmoid sinus, 1 point for the involvement of anterior and posterior periantral fat). The patient was given intravenous amphotericin to which the patient responded well. Repeat imaging was not done in this case.

Figure 3.

Severe disease with computed tomography severity index 18 (50-year-old male, coronavirus disease 2019 positive with periorbital pain, left eye vision loss and cranial nerve palsies). The patient was a known diabetic with an HbA1C of 12.2. Microbiological evaluation revealed Rhizopus species. A-C: Axial contrast-enhanced computed tomography (CT) images show bilateral disease involving bilateral maxillary sinuses and nasal cavities (asterisk in A) with involvement of right posterior antral fat and pterygopalatine fossa (PPF; arrow in A). There is soft tissue infiltration in the left orbital fat extending to the left inferior orbital fissure (IOF; asterisk in B) with evidence of thickening of the left globe wall along with tenting of the posterior pole of the globe giving a classical guitar pick appearance suggesting the involvement of the left globe (arrow in C); D-F: Coronal CT images show erosion of the hard palate (arrow in D), fat stranding in the left orbit and thickening and irregularity of the left optic nerve (arrow in E). There is evidence of soft tissue infiltration in the left cavernous sinus (dotted arrow in F) with focal narrowing of the cavernous segment of left internal carotid artery (ICA; arrow in F). The total summated computed tomography severity index, in this case, is 18 (1 point each for paranasal sinuses and nasal cavity involvement, 2 points for significant bilateral disease, 1 point for the involvement of posterior antral fat, 1 point for hard palate erosion, 1 point for orbital soft tissue involvement, 2 for PPF/IOF involvement, 3 points for globe involvement, 3 points for cavernous sinus involvement and 3 points for ICA narrowing). Left exenteration was done due to extensive disease and intravenous amphotericin was given. Clinically, the patient did not respond well to the treatment, and repeat imaging was done 10 weeks later.

Figure 4.

Repeat imaging after 10 weeks of patient in Figure 3 (known case of rhinoorbitocerebral mucormycosis with severe disease and post left exenteration). A and B: Axial contrast-enhanced computed tomography images (A and B) reveal soft tissue infiltrating the right posterior antral space (arrow in A) and soft tissue in the left orbit with contiguous extension to the left orbital apex and left cavernous sinus (arrow in B). Left post-exenteration status can be seen; C and D: Coronal images show soft tissue in the left cavernous sinus with the non-opacified cavernous segment of the left internal carotid artery suggesting its occlusion (arrow in C). There is also evidence of maxillary alveolus destruction and sequestrum formation (arrow in D). This repeat imaging shows radiological worsening post-treatment (non-responder).

Figure 5.

Receiver operating characteristic curve for computed tomography severity score and response assessment shows that the cut-off value for computed tomography severity index of 11 had a sensitivity of 78.26% and a specificity of 95.35% to predict response assessment. ROC: Receiver operating characteristic.

No significant association however could be established between history of diabetes or hypertension, type of causative organism or TALMI staging; and the response to treatment. The level of glycosylated hemoglobin also could not predict response to treatment (t-test, P = 0.25). However, the HbA1c values were significantly associated with CTSI values (t-test, P < 0.01) but with a weak correlation (Pearson coefficient 0.34). Age was not significantly associated with CTSI (t-test, P > 0.05). In addition, there was no agreement (kappa < 0.1) between the CTSI and TALMI staging.

DISCUSSION

Acute IFS is characterized by the invasion of neurovascular structures resulting in necrosis and spread beyond the sinonasal cavity. The presence of ophthalmologic or neurological complications carries a worse prognosis and can be potentially life-threatening. The two common causative organisms of IFS are from the Aspergillus species and Zygomycetes order[8,9].

COVID-19 infection causes immune dysregulation and predisposes the patients to other infections. The factors associated with an increased incidence of IFS in COVID-19 patients include uncontrolled diabetes mellitus, excessive use of corticosteroids for immunosuppression, and prolonged stays in the intensive care unit[10].

In the emergency setup, CECT PNS is the ideal modality for delineation of the extra sinus spread of disease to orbit, infratemporal fossa, skull base, and intracranial involvement along with the extent of bony erosions. An accurate delineation of the extension of disease on CT allows timely management with antifungal agents and surgical debridement.

We have proposed a scoring system on CECT (CTSI) to describe the severity of IFS in COVID-19 patients. This scoring system evaluates the disease extent to the sinonasal cavity, adjacent tissues, orbit, and intracranial disease on CT and further categorizes the disease into mild, moderate, and severe (Table 1). The previously available staging system of rhino-orbital-cerebral mucormycosis as described by Talmi et al[7] is primarily based on clinical evaluation. The critical areas of involvement including pterygopalatine fossa, superior and inferior orbital fissures, orbital apex, and infratemporal fossa have been given a higher value. Similarly, optic nerve and intraocular involvement in the orbit and cavernous sinus invasion, internal carotid artery narrowing, and intracranial complications were given a higher value.

In our study cohort, most patients had mild to moderate disease (Table 2) and TALMI stage 2 disease (46.97%). The mortality rate was 15.15% which is similar to 14% of COVID-19-associated rhino-orbital-cerebral mucormycosis (ROCM)[11] and less than 34% of coronavirus associated mucormycosis[12]. It has been suggested that though COVID-19 predisposes an individual to fungal infections, it does not appear to alter the prognosis in such patients. Also, in cases of ROCM, timely sinonasal debridement with appropriate antifungal therapy has been associated with higher survival rates. There was a significant association of bilateral disease with mortality. Bilateral sinus involvement increases the imminent risk to both the orbits[11] and intracranial disease and hence has a worse prognosis.

Fifty-seven patients (86%) had COVID before the diagnosis of IFS, ranging from a period of 1-36 days in comparison to the collaborative OPAI-IJO study on mucormycosis in COVID-19 study group[11] where 56% of cases had onset of ROCM within 14 days of COVID-19 diagnosis. In this study, all positive rapid antigen tests (RATs) were confirmed by a molecular test i.e., reverse transcription polymerase chain reaction (RT-PCR). It is essential to understand that COVID-19 test results can be influenced by multiple factors, such as the timing of sample collection, the type of test used, and the prevalence of the virus in the population. False positive results have been reported in 0.5% to 5% of cases. False positivity is of low concern in high prevalence situations. As this study was conducted during the COVID-19 pandemic when the prevalence was high, it is likely to be low. Potential causes of false positivity include mislabeling of samples during processing, cross-contamination of samples during collection and processing. RT-PCR result positive for a single gene should be viewed with suspicion and repeat testing should be done[13]. When using RAT, cross-reactivity of test antibodies with rheumatoid factor in patients with autoimmune diseases may also give a false positive result[14].

About three-fourths of the patients (n = 51) were known diabetic while a total of 56 patients had their glycosylated hemoglobin above 6.5 gm/dL (mean HbA1c of 9.11 ± 2.20 gm/dL). This incidence is similar to previous literature and it is believed that COVID-19 worsens the glucose profile of the patients with diabetes further predisposing them to mucormycosis[15]. Despite the associated co-morbidities, no significant association was found between survival and history of diabetes or hypertension, causative organism, CT severity score, or TALMI staging. The level of glycosylated hemoglobin also had no significant effect on mortality (t-test, P = 0.79).

Non-responders were defined as patients with residual clinical disease/ worsening and persistent disease in the follow-up scans (performed at 1-3 month intervals). Laterality and CTSI were significant predictors of response to treatment and CTSI of 11 had a sensitivity of 78.26% and a specificity of 95.35% to predict response assessment.

Out of the six patients who had bilateral disease, five were non-responders and bilaterality of disease had a significant association with poor clinical response as has been seen in previous studies. The treatment received by our cohort of patients included intravenous amphotericin B, surgery (functional endoscopic sinus surgery/paranasal sinus debridement/orbital exenteration in various combinations)[16]. Patients with a higher CTSI predominantly had orbital involvement in the form of intraocular or optic nerve involvement or intracranial disease. These sites especially disease at the orbital apex and cavernous sinus are not surgically resectable and show delayed response to medical management. Persistence of disease was seen at these sites in the majority of patients in the group of non-responders.

Collectively, these data indicated that the proposed radiological score is a useful guide in the diagnosis and follow-up of symptomatic patients with IFS. We believe that CTSI can help in the quantification of the disease burden and mapping out the extent of the disease for the surgeon. It can be useful in triaging patients at presentation and in response assessment during the hospital stay. A higher score on the initial CT would alert the clinician to initiate aggressive treatment, as severe disease correlates with slow response/non-response to treatment.

We encourage other institutions to test this scoring system and its correlation with the clinical status and response of patients of IFS (COVID-19 and non-COVID-19) to confirm its diagnostic efficacy.

The main limitation of our study was small sample size for comparison of radiological staging and clinical staging. Due to the limited sample size, comparison between mucormycosis and aspergillosis was also not possible in our study.

CONCLUSION

The CTSI is useful in quantification of the disease burden of IFS and response assessment in COVID-19 patients. It will also be useful in the triaging of patients at presentation, especially those with comorbidities like diabetes and hypertension.

Data sharing statement

Technical appendix, statistical code, and dataset available from the corresponding author at ashubhalla2@gmail.com. Participants gave informed consent for data sharing.