Journal of the Korean Rad iological Society, 1994 : 31 ( 1) : 63 - 67

Spiculation of lung Mass on CT: Carcinoma VS. Tuberculoma 1

Tae 11 Han, M.O. , Oong Wool‘ 5ung, M.O. ,

5eung Jae Lim, M.O. , YupYoon, M.O.

Purpose: 5piculation pathologically correlated with irregular fibrosis , loca­lized Iymphatic spread of tumor, or an infiltrative growth pattern of tumor, and the spiculation was ‘observed in malignant mass. But the spiculation was also observed in benign mass, particularly in tuberculoma. We retrospectively reviewed the length of spiculation under the hypothesis that the length of spiculation could be one of differential diagnostic points between lung cancer and tuberculoma.

Materials and Methods:We studied thirty seven patients (27 men and 10 women) ranging in age from 35 to 80 years (mean , 60 years). Analysis of spiculation included (a) the numberofspicules (b) the mean length ofspicul~s (c) the mean length of the longest spicule (비 the percentage of the mean length of spicules to the longest diameter of mass (e) the percentage of tþe mean length of spicules to the shortest dié! meter of mass.

Results: The mean length of spic비es of tuberculoma was 13.8 mm (5. 0.6.7) and that of lung cancer was 5.7 mm (5. D. 3.5). The percentage ofthe mean length of spic비es to the longest dia.meter of tuberculoma was 63.6% and that of lung cancer was 13.7%. The percentage of the mean length of spicules to the shortest diameter oftuberculoma was90.4% andthat of lung cancerwas 18.3% .

Conclusion: It is hard to differentiate lung cancer from tuberculoma on the basis of the spiculation being present or not, but we suggest that the longer spiculation be more highly suggestive of tuberculoma rather than lung cancer. The length of spiculation may help us differentiate lung cancer from tuberculoma.

Index Words: Lung , CT Tuberculosis, pulmonary Lung , Neoplasm

INTRODUCTION

In spite of many investigations for the CT appearance of lung -tumor interface and the internal texture of tumor , there st비 remains a significant difficulty in differentiating malignant lesions from benign lesions. Among the parameters for differentiating malignant lesions from benign lesions, spiculation is known to be pathologically correlated with i rregular fibrosis , localized Iymphatic spread of tumor, or infiltrative gro-

'Department ofD iagnostic Rad iology, Kyung Hee Uni vers ity Hospi tal Rece ived.November 4,1993 ; Accepted April 8, 1994 Addrèss reprint requests to :Tae 11 Han, M.D., Department 01 Diagnostic Radi. ology, Kyung Hee University Hospital. 1 Hoeki-dong, Dongdaemun.ku , Seoul , 130.702 Korea. Tel. 82.2.966.5191(2527) Fax. 82.2.968.0787

wth pattern of tumor, and it is a finding often observed in a malignant mass (1). However the spic비 ati on is al­so observed in benign mass such as tuberculoma ançl it is hard to differentiate lung cancer from tuberculo­ma on the basis ofthe spiculation bei [J g present or no1.

, Thepurpose of the this study is to differentiatelung cancer from tuberculoma by the length of the spiculation. We retrospectively reviewed the length of spiculation under the hypothesis tnat the length of spiculation could be a differe r1 tial diagnostic point be­tween lung cancer and tuberculoma.

MATERIALS and METHODS

CT scans of 250 patients performed during Septem­ber, 1989 to August, 1992 were reviewed for pulmonany

n

ω

Journal of the Korean Radiologica l Society, 1994 ; 31 ( 1) : 63-67

nodules without the knowledge on the confirmed diag­nosls

We defined spiculation as linear parenchymal strands radiating from peripheral margin of the lung mass into the surrounding lung parenchyma

We excluded the spiculations with (a) branching pat­tern , (b) pleural tag , or (c) faint density (Fig. 1). Branching pattern of spic비 ation was excluded be­cause it was difficult to differentiate from p비 monary vesse l. Pleural attachment of the spic비 ation was excluded because it was difficult to measure the length ofspiculation precisely.

Thirty seven patients (27 men and 10 women) ran­ging in age from 35 to 80 years (mean , 60 years) ful­filled the above criteria. Cytopathologic confirmation could be obtained through bronchoscopy (9 patients) , percutaneous needle aspiration (12 patients) , sputum cytology (6 patients) and operation (2 patientsl. Eight cases were diagnosed to be tuberculoma which showed no sign of increase in size during the 1 year fol­low - up period.

CT scans were obtained with GE 9800 Quick scanner. AII scans were performed at 1 cm intervals from the apeces to the lung base with the use of intravenously administered contrast materia l. Technical scan parameters were 120 kVp , 120 mA, 2-second scan time , 10mm scan thickness , 40 -cm field of view, and standard reconstruction algorithm. AII scans were photographed at window and level setting appropr iate for mediatinum (Ievel= -15 - 45HU , width = 400-450 HU) and lung parenchyma (Ievel = -600 - -850HU, width 850 - 1500HU)

On the sl ice showing the largest diameter of the lung

a

mass the spi culations were evaluated in regard to (a) the number of spicules , (b) the mean length of spicules , (c) the mean length of the longest spi cule , (d) the per­centage of the mean length of spicules to th e longest di­ameter of mass , and (e) the percentage of the mean length of spic비 es to the shortest diameter of mass. Since the diameter of the mass influenced the length of spiculation , we obtained the percentage of length of spicules to mass. Parametric data were analyzed using the Student t test , and statistically si gn ificant difference was accepted when P was .05 or less.

RESU L TS (Table.1)

Among the 37 patients , 23 patients (62 %) had lung cancer and 14 patients (38 %) had tuberculoma

The mean number of spic비 es of lung cancer (mean ::!:: standard deviation , 3.1 ::!:: 1.9) was not significantly greater (p > 0.01 ; 95 % confidence i ntervall than that of tuberculoma(2.1 ::!:: 1.3).

The mean length of spic비 es of tuberculoma (mean ::!:: S. 0 ., 13.8 mm ::!:: 6.7 mm) was significantly greater (p < 0.01) than that of lung cancer (5.7 mm ::!:: 3.5 mm). There were 14 cases with the spic비es more than 10 mm in length , of which 11 cases were tuberculoma (Fig. 2a) In most cases of lung cancer , the length of spic비 eswas under 10 mm. The mean length ofthe longest spicule of tuberculoma (mean ::!:: S.O., 15.6 mm ::!:: 7.5 mm) was significantly greater (p < 0.01) than that of lung cancer (7.0mm ::!:: 4.0mm).

The percentage of the mean length of spicules to the longest diameter of tuberculoma (mean ::!:: S. 0. , 63.6 % ::!:: 40.7 %) was significantly greater (p < 0.01) than that

c

Fig. 1 . Spiculati on excluded in thi s study : (a)branching pattern (b)pleural tag (c)fain t density y‘ - 64 -

Table 1 . Analysis of spiculation

Mean length of spic비es

Mean length of the longest spic비e

Mean length of spicules

The longest diameter of mass

Mean length ofthe longest spicules

The longest diameter of mass

Mean length of spicules

The shortest dimater of mass

Mean length ofthe longest spic비 es

The shortest diameter of mass

Number of spic미 es

Lung Tuber- Analysis cancer culoma

5.7mm 13.8mm p ( 0.01

7.0mm 15.6mm p ( 0.01

13.7% 63.6% P ( 0.01

17.3% 69.3% P ( 0.01

1 8.3 % 90.4 % P ( 0.01

23.1 % 98.6 % P ( 0.01

3.1 2.1 p ) 0.01

Tae 11 Han , et al ‘ Spiculation of Lung Mass on CT

。f lung cancer (13.7% :t 8.6%). The relative ratio of the mean length of spicules to mass diameter was above 25% in most cases of tuberculoma and it was below 25 % in lung cancer (Fig. 2b). The percentage of the mean length of spic비 es to the shortest diameter of tuberculoma (mean :t S.O. , 90.4% :t 64.2%) was sig­nificantly greater (p < 0.01) than that of lung cancer (18.3 % :t 15.2%)

DISCUSSION

Spiculation has been known as the characteristic finding of lung cancer in the past, but it has been fre­quently reported also in tuberculoma. In our cases , spiculation was identified in 14 cases of tuberculoma and it was not easy to differentiate lung cancer. from tuberculoma on the basis of the spiculation being pres­ent or not. However, we could differentiate lung cancer

MEAN SPICULA TION LENGTH Mean length of Spiculation/Longest Diameter of the Mass

30

" 20

mm

" 10

100

90

80

70

60 M , .. 50

40

30

20

10

‘ 6 7 B 9 10 11 ’ 2 13 ’‘ ’ 5 16 17 18 19 20 21 22 23 3 4 5 9 10 11 12 13 1‘ 15 16 17 18 19 20 21 22 23

No‘ 。f Patients

a b Fig. 2. a. The mean length of spicules b. The percentage ofthe mean length of spicules to the longest diameter oftuberculoma

/

ML= 6.7 mm a b

- 65 -

NO. of Patients

Fig . 3. Long spicules in tuberculoma, whe­reas short spicules in lung cancer : a. In the tuberculoma, mean length of spi-cules is 27 mm‘

b. ln the lung cancer , 6.7 mm.

Journal of the Korean Radiological Society, 1994: 31 ( 1) : 63 - 67

a

a

.... ML= 8mm ML/LO= 20%

b

i톰 V

b

from tuberculoma by the spiculation length and the relative ratio (of the spicule length to the mass diam­eterl. The spiculation length (Fig. 3) and the relative ratio (Fig. 4) of tuberculoma were significantly greater than that of lung cancer

There are many published reports (1-7) on differen­tiation of benign from malignant lesions , and most of these suggested following criteria (2) for favoring a di­agnosis of malignant tumor: (a) an intrap비 monary

nodule with irregular , fuzzy borders , finger-like projections , or fine spiculations (b) a diameter of 4 cm 。 r larger (c) a CT number of approximately 50 HU (45 to 65 HU) and (d) absence of CT evidence of calcification But there was considerable overlap in above criteria between benign and malignant lesions. Because ofthe fact that many exceptions can be observed in each indi­vidual criterion , the diagnosis of probable malignant lesion can be made with considering many criteria Spiculation was observed in seventy two out of the eighty two primary carcinomas and two out of the three metastases, but was also present in four out of the five tuberculoma lesions and in one out of the two inflam­matory pseudotumor by Charles (1). In Keiko Kuriyama’s report (8) , spic비 ation was seen in 78% of small peripheral lung cancers. AII of the papillary adenocarcinomas had spiculations. The spiculated margins were due to the tumor invasion of the sur­rounding lung tissue. Stanley (9) observed the edge analysis that 52 out of 66 nodules with sharp and smooth margin were benign , but 14 (21 .2%) out of 66 nodules were malignan t. Although there was consider­able overlap in edge characteristics (10) of each type of lesion , the majority of primary bronchogenic carci­noma had spiculated margins , whereas only a minority

- 66

Fig . 4. Higher percentage ratio (mean length 01 spicules/the longest diameter 01 mass) in tuberculoma, whereas lower per­centage ratio in lung cancer : a. In the tuberculoma, percentage (M LlLO =23 %, MLlSO=53 %) b. In the lung cancer , (M LlLO=11 %, MLI SO =13 %). (ML : mean length 01 spicules,

LO : the longest diameter 01 mass , SO: the shortest diameter 01 mass)

Fig. 5. The distinction 01 a tuberculoma lrom lung cancer cannot be made only by the length 01 spicules a. In the tuberculoma, mean length 01 spi­cules is 8 mm and MLlLO is 20%. b. In the lung cancer , mean length 01 spi­cules is 10 mm and MLlLO is 32%

。f carcinoid tumors, metastatic nodules, and benign lesions did so. The shaggy margins of malignant neo­plasm are related to the so-called corona radiata (11). This sign consists of a number of fine , linear striations that extend perpendic비 arly outward from the periph­ery of a nodule for a distance of perhaps 4 or 5 mm. While the corona radiata is not diagnostic of malig­nancy (also seen in the lesions of progressive massive fibrosis of silicosis and in lipid pneumonia) , it constitutes a highly suggestive sign that should be regarded with a high index of suspicion. However, as pointed out by Heitzmann (12) , its presence does not constitute a pathognom ic sign of malignancy

Pathologically (1) , spiculation most commonly is correlated with a desmoplastic response in the nodule , resulting in fibrotic strands radiating into the sur­rounding lung parenchyma. Among malignant leisons, SplC비 ation is occasionally associated with direct infil­tration of the tumor into adjacent bronchovascular sheaths or localized Iymphangitic extension. Unfor­funately , there is not distinguishable from spiculation due to a fibrotic response. Among benign lesions, spi­cuation is due exclusively to a desmoplastic response.

There has not been any published reports on spi­culation length. Although the exact cause of the longer spiculation in tuberculoma is not known , authors hypothesize that the tuberculoma is associated with longer stranded parenchymal scarring

Although we have found that the longer spiculation is a suggestive finding of tuberculoma rather than lung cancer , the distinction of a tuberucloma from lung can­cer cannot be made only by the spiculation length (Fig. 5).

Conclusively it is hard to differentiate lung cancer

from tuberculoma on the basis of the spiculation being

present or not, but we suggest that the longer spicu­

lations be highly suggestive oftuberculoma rather than

cancer. The length of spiculation may help us differen­

tiate lung cancer from tuberculoma

REFERENCES

1. Zwirewich CV, Vedal S, Miller RR, Muller NL. Solitary Pulmonary

nodule. High-Resolution CT and radiologic pathologic corre­

lation. Radiology 1991 ; 179: 469-476

2. Shin MS, Ho KJ. Computed tomographic evaluation 01 solitary

p비 omonary nodules in chest roentgenograms. J Comput Assist

Tomogr1982 ; 6: 947-954

3. Khouri NF , Meziane MA, Zerhouni EA, Fishman EK , Siegelman

SS. The solitary pulmonary nodule: assessment, diagnosis , and

management. Chest1987; 91 : 128-133

4. Theros EG. Varying manilestation 01 peripheral pulmonary neo­

plasm: A radiologic-pathologic correlative study. AJR 1977; 128

893-914

Tae 11 Han , et a/ ‘ Spiculation of Lung Mass on CT

5. Kim KY, Choe KO. Analysis 01 computed tomographic mani­

lestations 01 primary lung cancer by histologic types. Journalof

Korean Radiological Society 1988 ; 24 : 1 025-1 034

6. Kim H, Kim OB , Woo SK, Suh SJ. The computed tomographic

lindings 01 bronchogen ic carcinoma presenting as a solitary per­

ipheral pulmonary mass. Journal of Korean Radiological Society

1985; 21 ‘ 719-726

7. Zwirewich CV‘ Miller RR , Muller NL. Multicentric adenocarci­

noma 01 the I ung : CT-path이 ogic correl ation. Radiology 1990 ;

176 : 185-190

8. Kuriyama K, Tateishi R, Doi 0 , et al. CT-pathologic correlation in

small peripherallung cancers. AJR 1987; 149: 1139-1143

9. Zerhouni EA, Stitik FP, Siegelman SS, et al. CT 01 the p비 monary

nodule : A cooperative study. Radio/ogy 1986 ; 160 : 319-327

10. Siegelman SS‘ Khouri NF, Leo FP, Fishman EK , Braverman RM , Zerhouni EA. Solitary p비moanry nodules: CT assessment. Radi­

ology 1986 : 160 ‘ 307-312

11 . Fraser RG , Pare JAP, Pare PD , Fraser RS, George P. Genereux

GP. Diagnosis of disease of the chest. 3rd ed. Philadelphia

Sauders , 1989 ; 1339-1407

12. Heitzman ER. The lung: radiologic-pathologic correlation. 2nd

ed. StLouis: Mosby, 1984 ; 384-396

대한방사선의학회지 1994; 31( 1) : 63-67

폐종괴의 침상화:암종 대 결핵종

경희대학교 의과대학 진 단방사선과학교실

한태일·성동욱·임승재·윤 엽

목 적.흉부 전산화 단층촬영상 (이하 CT) 폐종괴 주변부에 보이는 spiculation은 폐암이 폐조직에 직접 침윤 혹은 결합

조직 형성 반응에 의한 것으로， 과거 폐암의 특징적인 소견으로 생각했으나 앙성종앙에서도 종종 관찰된다. CT상 spiculation

을 관찰함으로써 spiculation의 길이가 앙성종앙과 악성종양을 감별할 수 있는 인자가 될 수 있다는 가정하에 본 연구를 시행

하였다

대상 및 방법 CT를 시행하고 조직학적 및 임상적으로 확진되고 sp iculation을 가진 폐종괴 37예를 대상으로 하였다.

Spiculation의 관찰은 lung setting상 종괴의 직경이 가장 긴 부위에서 (j) spic미es의 수 (g) spicules의 평균 길이 @ 가장 긴

spicule 길이의 평균 @ 종괴의 최대직경에 대한 spicules의 평균길이의 비율 @종괴의 최소직경에 대한 spicules의 평균길

이의 비율을 측정하여 통계학적으로 분석하였다.

결 과 : 전체 37예중 폐암이 23예 ( 62%) , 결핵걸절이 14예 (38%) 였다. 걸핵결절의 경우 spicules의 평균길이는 13.8 mm

(SD 6.7) 였고， 폐암은 5.7 mm (SD 3.5) 였으며 제일 긴 spicule의 길이의 평균은 결핵걸절에서 15.6 mm (SD 7.5) 였고 폐암

은 7.0 mm (SD 4.0 ) 였다. 종괴의 최대 직경에 대한 spicules의 평균 길이의 비는 결핵결절이 63_6% . 폐암이 13_7%였고， 종괴

의 최소직경에 대한 비는 걸핵결절이 90.4%, 페암이 18.3%였다. CT상 한 단면에서 spicules의 갯수는 평균적으로 결핵결절

2.1개， 폐암 3.17H였다.

결 론 Spiculation 으| 유무만으로는 악성과 양성을 감멸하기 어려우나 spiculation의 길이가 길면 폐암보다는 결핵결절을

시사한다. 따라서 SplC비ation의 길이는 결핵결절과 폐암을 감별하는데 도움이 될 것으로 사료된다.

- 67 -

’ 94년도 대한방사선의학회 제50차 학술대회 초록제출 양식

저| 도--a， .

저 자:

-ι4ι- i-1: -•

구연 자:

구 분: 口구 연

口 전 시 Space 개 (76cm x lOOcm)

口 Back Board Type 디 Illuminating Type

口 뇌신경계 부 ’ 겨

。

두

口 口 심혈관계 口 흉부 口소화기계

口 버뇨생식계 口근골격계 口 소아 口 핵의학 口유방 口 기타

드2. ^~ . ..., -, .

대상및 방법 :

결 과:

결 론:

- 68 -