If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address for correspondence: Nobuhiko Seki, MD, Division of Medical Oncology, Department of Internal Medicine, Teikyo University School of Medicine, Itabashi-ku, Tokyo 173-8605, Japan
Studies of lung cancer showing localized ground-glass attenuation (GGA) on thin-section computed tomography (TSCT) have been limited to resected stage IA adenocarcinomas. This study aimed to clarify the features of localized GGA cancer as a distinct clinicoradiological entity through a survey of lung cancers of all types.
Methods
From 2000 through 2002, 492 patients with newly diagnosed stage I–IV lung cancer underwent TSCT at a single institution. The tumors were semiquantitatively classified into four groups on the basis of GGA area as a percentage of the whole tumor shadow (GGA ratio) on TSCT images: 100%, 99–50%, 49–1%, and 0%. The relationship between clinicopathological data and the GGA ratio, predictors of the presence of GGA, survival data, and prognostic factors were evaluated retrospectively.
Results
All localized GGA cancers were adenocarcinomas (p < 0.05). A GGA component was not found in patients with advanced cancer (p < 0.05). GGA cancer was related to nonsmoking status (Odds ratio 6.17, p < 0.05). A threshold tumor size of 30 mm in GGA cancer (hazard ratio, 2.86; p < 0.01) and the GGA ratio (hazard ratio, 4.17; p < 0.01) were independent prognostic factors. Survival rates were higher in patients with a GGA ratio ≥50% and stage IB lung cancer than in patients with a GGA ratio <50% and stage IA lung cancer.
Conclusion
Localized GGA cancer, with presurgical prognostic factors of tumor size and GGA ratio, represents early-stage lung adenocarcinoma in nonsmokers.
Localized ground-glass attenuation (GGA) has become a major concern as lung cancer screening with low-dose helical computed tomography (CT) becomes more widely available.
Although GGA is usually a nonspecific finding associated with several pulmonary disorders, some recent studies suggest that localized, well-circumscribed areas of GGA often represent malignant tumors.
Focal area of ground-glass opacity and ground-glass opacity predominance on thin-section CT: discrimination between neoplastic and non-neoplastic lesions.
Other recent reports have shown that such tumors are often adenocarcinomas, in particular, bronchioloalveolar carcinoma (BAC) or adenocarcinoma with a BAC component,
Proportion of ground-glass opacity on high-resolution computed tomography in clinical T1 N0 M0 adenocarcinoma of the lung: a predictor of lymph node metastasis.
Therefore, lung cancer showing a GGA ratio of 100–1% on TSCT (localized GGA cancer) might be recognized as a clinicoradiological entity distinct from conventional lung cancer showing homogeneous consolidation of a nodule or mass (non-GGA cancer).
However, most studies of localized GGA have included only surgically resected adenocarcinomas no larger than 30 mm in diameter.
Focal area of ground-glass opacity and ground-glass opacity predominance on thin-section CT: discrimination between neoplastic and non-neoplastic lesions.
Proportion of ground-glass opacity on high-resolution computed tomography in clinical T1 N0 M0 adenocarcinoma of the lung: a predictor of lymph node metastasis.
Such studies cannot document all features of localized GGA cancer. Therefore, to establish localized GGA cancer as a distinct clinicoradiological entity, studies should include lung cancers of all types, including adenocarcinomas 30 mm or greater in diameter and stage II–IV carcinomas. Moreover, localized GGA cancers should be compared with non-GGA cancers by means of demographic and survival data.
Findings to date raise several questions. (1) Would all localized GGA cancers be adenocarcinomas if all types of lung cancer were evaluated? (2) Is a GGA component present in patients with advanced (inoperable clinical stage IIIB or IV) lung cancer? (3) Is GGA cancer related to age, sex, or smoking history? (4) Is the tumor size of GGA cancer, in addition to the GGA ratio, an independent prognostic factor? (5) If so, is the prognosis better for patients with a higher GGA ratio but larger tumors or for patients with smaller tumors and a lower GGA ratio?
This study aimed to clarify the features of localized GGA cancer as a distinct clinicoradiological entity by addressing these questions through a survey of lung cancers of all types.
PATIENTS AND METHODS
Patients
We reviewed the records of 529 consecutive patients with primary lung cancer newly diagnosed at the National Hospital Organization Shikoku Cancer Center from June 2000 through May 2002. Although TSCT was performed routinely, scans were not available for 37 patients; therefore, 492 patients were evaluated. The tumor stage was established with the International Staging System.
Tumors were stage I in 237 patients, and stage II–IV in 255 patients. In cases of synchronous lung cancer, the most advanced tumor was considered the index tumor.
Our institutional review board does not require its approval or patient informed consent for retrospective studies of case records and CT images.
Determination of GGA Ratio on TSCT Images
Patients underwent TSCT without contrast enhancement using a Somatom Plus 4 scanner (Siemens Medical Systems, Erlangen, Germany; 120 kVp, 250 mA, 1.0-second scanning time, 2.0-mm-thick sections). Localized GGA was defined as a focal area of GGA with smooth and distinct borders on TSCT images with a fixed lung window setting (a level of −700 and a width of 1500 Hounsfield units). The GGA ratio was calculated semiquantitatively as [(DGGA − D)/DGGA] × 100, where DGGA and D are the greatest diameters of the entire tumor and the non-GGA component, respectively.
Tumors were divided on the basis of GGA ratio into four groups: 100%, 99–50%, 49–1%, and 0% (Figure 1). Tumors were assigned to groups according to the consensus judgment of two investigators (N.S. and K.E.), who were blinded to clinical information.
FIGURE 1Transverse thin-section computed tomography images of lung adenocarcinoma. Tumors were classified into four groups according to semiquantitatively calculated GGA ratios. (A) 100% GGA ratio tumor, (B) 99–50% GGA ratio tumor, (C) 49–1% GGA ratio tumor, and (D) 0% GGA ratio tumor.
Pathology and Genetics of Tumours of the Lung, Pleura, Thymus and Heart (World Health Organization Classification of Tumours). IARC press,
Lyon2004: 9-124
by a lung pathologist (R.N.), who reviewed all specimens for this study. The histologic diagnosis was adenocarcinoma in 308 patients, squamous cell carcinoma in 108 patients, small cell carcinoma in 51 patients, large cell carcinoma in 12 patients, and other non-small cell carcinoma in 13 patients. Adenocarcinoma was subclassified as BAC or adenocarcinoma with or without a BAC component.
Tumor Resection and Follow-up
The patient flow of this study is summarized in Figure 2. Of the 237 patients with clinical stage I lung cancers, 225 underwent surgery, two received chemoradiotherapy, six received radiotherapy, and four received best supportive care. Patients undergoing resection of stage I lung cancer received neither radiotherapy nor chemotherapy before or after surgery. From June 2000 through December 2001, thoracoscopic wedge resection was performed for tumors 10 mm or less in diameter with a GGA ratio of 100%.
Although this is not a prospective study and the follow-up procedure was, therefore, not strictly regimented, follow-up care by the same two surgeons (S.S. and M.N.) at a single institution allowed patients to be followed up with standardized examinations. Physical examination, chest roentgenography, and tumor markers as appropriate were performed every month within the first year and at 3- to 6-month intervals thereafter. Chest CT, abdominal ultrasonography, brain CT or magnetic resonance, and bone scintigraphy were performed every 6 months through the third year and annually thereafter. If a relapse was detected with these imaging techniques, the patient underwent further examinations as required. All patients were followed up, for a median duration of 36 months (range, 1–78 months).
Statistical Analysis
To compare data between groups, one-way analysis of variance or the Kruskal-Wallis test was used, followed by Scheffe's test. Spearman's rank correlation coefficients were analyzed for relationships between groups. Multivariate logistic regression analysis was performed for predictors of GGA. Relapse-free survival curves were constructed with the Kaplan-Meier method, followed by the log-rank test. Multivariate Cox proportional hazards model analysis was performed for prognostic factors. All calculations were performed with Stat View 5.0J software (SAS Institute Inc., Cary, NC). p values <0.05 indicated statistical significance.
RESULTS
Relationship Between Clinicopathological Data and GGA Ratio
Patient's clinicopathological data according to GGA ratio are summarized in Table 1. Age was lowest in the 100% GGA ratio group (p < 0.05) but did not differ significantly between the other three groups. The percentages of female patients and of nonsmokers were significantly higher (p < 0.05, p < 0.05, respectively) in groups with GGA ratios of 100%, 99–50%, or 49–1% than in the group with a GGA ratio of 0% but did not differ significantly among the former three patient groups. Similarly, the frequencies of clinical stage I and of adenocarcinoma were significantly higher (p < 0.05, p < 0.05, respectively) in these three groups than in the 0% GGA ratio group. Neither clinical stage II–IV disease nor cancers other than adenocarcinoma were found in these three groups. Tumors were smallest in the 100% GGA ratio group and largest in the 0% GGA ratio group (p < 0.05).
TABLE 1Patients' Clinicopathological Characteristics According to GGA Ratio
Multivariate logistic regression analysis demonstrated that only smoking history (p < 0.01) was significantly associated with GGA (Table 2). The likelihood of GGA in nonsmokers was 6.17 times that in smokers.
Relationship Between Histopathological Data and GGA Ratio
Pathologic characteristics of resected clinical stage I lung cancer according to GGA ratio are summarized in Table 3. The rate of clinical stage IA disease was highest in the 100% GGA ratio group (p < 0.05) and did not differ significantly between the other three groups. Clinical stage IB disease was not observed in the 100% GGA ratio group. The rate of pathologic stage IA disease was highest in the 100% GGA ratio group and lowest in the 0% GGA ratio group (p < 0.05). Neither pathologic stage IB nor II-IIIA disease was found in the 100% GGA ratio group, whereas neither pathologic stage IIIB nor IV disease was observed in any of the four groups. Furthermore, pathologic stage was negatively correlated with the GGA ratio (r = −0.33, p < 0.001). Similarly, the rate of BAC was highest in the 100% GGA ratio group and lowest in the 0% GGA ratio group (p < 0.05). Adenocarcinoma without a BAC component was not observed in the 100% GGA ratio group, whereas cancers other than adenocarcinoma were found only in the 0% GGA ratio group. Moreover, histologic subtype was negatively correlated with the GGA ratio (r = −0.35, p < 0.001).
TABLE 3Pathological Characteristics of Resected Clinical Stage I Lung Cancer
After independent variables were selected from among presurgical data, multivariate analysis with the Cox proportional hazards model revealed that clinical stage IA or IB disease (p < 0.01) and the GGA ratio (p < 0.01) were independent risk factors for relapse-free survival in 225 patients with clinical stage I lung cancers (Table 4). The likelihood of relapse in patients with clinical stage IB disease was 2.86 times that in patients with clinical stage IA disease. Furthermore, the risk of relapse increased 4.17-fold with the stepwise decrease of the GGA ratio from 100%, through 99–50% and 49–1%, to 0%.
When a tumor size of 20 mm was used instead of 30 mm as a cut-off value, the GGA ratio was the only significant independent risk factor.
Combined Effects of GGA Ratio and Tumor Size
Relapse-free survival rates, stratified by both the GGA ratio and clinical stage, are shown in Figure 3. The 5-year relapse-free survival rate, especially that of patients with stage IB disease, was highest (100%) when the GGA ratio was ≥50% (n = 10), whereas that of patients with stage IA disease was 95.5% when the GGA ratio was 49–1% (n = 35) and was 67.7% when the GGA ratio was 0% (n = 81). Relapse occurred through local recurrence or distant metastasis.
FIGURE 3Relapse-free survival curves stratified by both GGA ratio and clinical stage. The 5-year relapse-free survival rate of patients with stage IA disease (n = 55) or stage IB disease (n = 10) was highest (100%) when the GGA ratio was ≥50%. When the GGA ratio was 49–1%, the 5-year relapse-free survival rate of patients with stage IA disease (n = 35; 95.5%) was significantly greater than that of patients with stage IB disease (n = 14; 61.9%, p = 0.047). Similarly, when the GGA ratio was 0%, the 5-year relapse-free survival rate of patients with stage IA disease (n = 81; 67.7%) was significantly greater than that of patients with stage IB disease (n = 30; 54.6%, p = 0.019).
The present study has demonstrated that clinical features differ between localized GGA cancer and non-GGA cancer in the following ways: (1) The histologic type of localized GGA cancer is limited to adenocarcinoma. (2) Localized GGA cancer is not present in patients with advanced lung cancer. (3) Localized GGA cancer is related to nonsmoking status. (4) The size of localized GGA cancer (≤30 mm or >30 mm) is, in addition to the GGA ratio, an independent prognostic factor, as it is for non-GGA cancer. (5) Patients with a GGA ratio ≥50% and stage IB lung cancer have a better prognosis than do patients with a GGA ratio <50% and stage IA lung cancer. To our knowledge, this is the first study to clarify the features of localized GGA cancer as a distinct clinicoradiological entity.
In the present study, all localized GGA cancers were adenocarcinomas, although lung cancers of all types were examined. To date, most studies of localized GGA cancer have cited four earlier studies as evidence that localized GGA on TSCT images indicates focal BAC, mostly representing a growth pattern of tumor cells replacing alveolar lining cells.
Focal area of ground-glass opacity and ground-glass opacity predominance on thin-section CT: discrimination between neoplastic and non-neoplastic lesions.
Kushihashi et al. have correlated only four bronchioloalveolar adenomas, but not BAC, with TSCT images from among 668 lung cancers but did not mention the presence of GGA among other cancers.
Moreover, Sakai et al. have reported that no localized GGA was observed on TSCT images in 30 cases of surgically resected peripheral squamous cell carcinomas, including eight carcinomas with alveolar lining tumor permeation and fibrous thickening of alveolar septa,
which might pathologically mimic adenocarcinoma with a BAC component and corresponds to the diffuse desmoplastic type of squamous cell carcinoma proposed by Tokuda et al.
Therefore, ours is the first study providing evidence that localized GGA cancers are adenocarcinomas, especially BAC or adenocarcinoma with a BAC component.
Furthermore, all localized GGA cancers in our study were clinical stage I and never progressed to pathologic stage IIIB or IV, although all types of lung cancers were evaluated. To date, most studies of localized GGA cancer have involved only resected clinical stage IA adenocarcinomas, which do not exceed pathologic stage IIIA.
Proportion of ground-glass opacity on high-resolution computed tomography in clinical T1 N0 M0 adenocarcinoma of the lung: a predictor of lymph node metastasis.
However, we wonder whether localized GGA cancer of any size is present in patients with inoperable clinical stage IIIB or IV disease. This question is difficult to answer on the basis of previous reports, considering that TSCT is not routinely performed for clinical stage IIIB or IV primary lung cancers. In our study 176 of 195 patients with clinical stage IIIB or IV primary lung cancer underwent TSCT, but no localized GGA cancer was observed. An important question is why localized GGA cancer was not found in patients with clinical stage IIIB or IV disease, although metastasis to mediastinal lymph nodes had occurred before surgery in some cases and postoperative relapse or distant metastasis was possible in other cases. Although the reason localized GGA cancer was not found remains unclear, some clues have been obtained. Saito et al. have reported that, during their long-term follow-up of eight localized GGA cancers, vascular involvement increased in all cases, areas of high attenuation (non-GGA component) appeared within the areas of GGA before an increase in the size of the entire tumor in seven cases, and a marked increase in the size of the non-GGA component was observed in two cases.
Therefore, our proposed explanation is that the progression of GGA cancer might be followed by both increased metastatic potential and architectural distortion of the GGA component.
That localized GGA cancer is a distinct clinicoradiological entity is supported by differences in time-independent factors (gender, smoking history, and histologic type; Table 1) between the 0% GGA ratio group and the groups with GGA ratios of 100%, 99–50%, and 49–1% and by the lack of differences among these latter three groups. Furthermore, multivariate analysis showed that localized GGA cancer was related to nonsmoking status but not to the female sex. To date, anecdotal evidence has suggested that localized GGA cancer tends to be detected in female nonsmokers. If localized GGA cancer represents, in particular, BAC or adenocarcinoma with a BAC component, the percentage of nonsmokers correlates with the percentage of female subjects, as in epidemiologic studies of BAC.
Moreover, Kosaka et al. have shown with univariate analysis that epidermal growth factor receptor mutations, which are common in BAC and adenocarcinoma with a BAC component, are more frequent in woman and in nonsmokers, but Kosaka et al. have also shown with multivariate analysis that epidermal growth factor receptor mutations are independently associated with nonsmoking status but not with the female sex.
In the present study, some adenocarcinomas without BAC component showed GGA on TSCT (Table 3). The histologic subtype of all these adenocarcinomas was papillary adenocarcinoma. Regarding the specific findings of these tumors, the area of air space in pathologic specimens was large because of a relative absence of thick papillary structures, compact cellular growth pattern, and acinar filling. Therefore, considering that the presence of localized GGA is associated with a high percentage of air-containing areas in the tumor, it seems reasonable that some papillary adenocarcinomas with such specific findings showed localized GGA on TSCT and that the other subtypes of lung carcinoma presenting a solid and nesting growth pattern with dense fibrous tissue showed a GGA ratio of 0% on TSCT. In fact, Aoki et al. have demonstrated that 3.2% of adenocarcinomas without a BAC component showed a GGA ratio of 10% to 50% on TSCT.
However, more important is the fact that the prognosis of localized GGA cancers, even if adenocarcinomas without a BAC component are included, is better than that of non-GGA cancers. This more-favorable prognosis is the reason localized GGA cancer is an important clinicoradiological entity.
Our study found that size is an important factor in localized GGA cancer as well. That is, for patients with clinical stage I disease, both tumor size and the GGA ratio might be independent prognostic factors when a threshold value of 30 mm is used. In previous studies, a tumor diameter of less than 30 mm was found not to be an independent prognostic factor for localized GGA cancer.
Proportion of ground-glass opacity on high-resolution computed tomography in clinical T1 N0 M0 adenocarcinoma of the lung: a predictor of lymph node metastasis.
Therefore, our finding that a tumor size of 30 mm has prognostic value, despite our other finding that a tumor size of 20 mm has no prognostic significance, strengthens the reliability of our results.
The results of our study of the combined effects of the GGA ratio and tumor size in patients with clinical stage I disease and localized GGA cancer have several interesting implications. First, in patients who had tumors with a GGA ratio ≥50%, the 5-year relapse-free rate was 100% for both clinical stage IA and IB disease. Therefore, the clinical stage of patients who have tumors with a GGA ratio ≥50% might need to be decreased below the conventional clinical stage of IA, regardless of tumor size. Second, patients with a GGA ratio ≥50% and stage IB lung cancer have a better prognosis than do patients with a GGA ratio <50% and stage IA lung cancer. In two recently published series, the frequency of multicentric adenocarcinoma was 8.4 to 21.5%,
So far, the most common surgical procedure for such cases is a combination of lobectomy for the most-advanced tumors and limited resection for less-advanced tumors.
Therefore, our results might be useful for selecting treatment for patients with clinical stage IA or IB disease and synchronous GGA cancer.
Our study had several limitations. A first limitation is that although analysis with CT software for calculating the GGA ratio 2- or 3-dimensionally might have been more accurate, it was not available for our study. However, in most previous studies, semiquantitative methods have been employed to calculate the GGA ratio, resulting in sufficient clinical benefit.
Proportion of ground-glass opacity on high-resolution computed tomography in clinical T1 N0 M0 adenocarcinoma of the lung: a predictor of lymph node metastasis.
A second limitation is that our study is a retrospective cohort study of patients from a single cancer-center hospital, which may have led to a selection bias of patients with localized GGA cancers. If larger prospective cohort studies are performed, localized GGA cancer other than early-stage adenocarcinoma might be detected. Therefore, the possibility of such bias might limit the strength of our conclusions. In fact, advanced adenocarcinomas, such as mucinous adenocarcinomas, can manifest as solitary or multifocal GGA. Moreover, lung cancers of other histologic types can manifest as localized GGA due to peritumoral hemorrhage regardless of stage. However, previous reports regarding mucinous and nonmucinous BAC have indicated that advanced mucinous adenocarcinomas very rarely show solitary GGA and that most show the combination of multifocal GGA and non-GGA lesions.
In addition, Li et al. have demonstrated that persistent localized GGA is most frequently attributed to BAC and that short-term follow-up imaging can be helpful for differentiating benign from malignant GGA lesions, because all benign GGA lesions in their series, including those with peritumoral hemorrhage, had partially or completely resolved within 3 months.
A third limitation is that regarding the strict diagnosis of adenocarcinoma with multiple localized GGA, the distinction between multicentric origins (field cancerization) and metastasis remains poorly defined. Therefore, in our study, whether adenocarcinoma manifested as multiple localized GGA can be considered stage I lung cancer was decided on the basis of clinical and pathologic criteria.
In the future, a better understanding of the molecular nature of such cancers may open new avenues for the study of patients with multiple localized GGA. A fourth limitation of the present study is that patients with metastatic lung tumors were not evaluated. In the setting of neoplastic diseases other than lung cancer, localized GGA on TSCT is considered to reflect lung metastasis from adenocarcinoma of the gastrointestinal tract or pulmonary involvement of lymphoma or malignant melanoma.
However, Steele has shown that only 3% of 887 asymptomatic patients with solitary pulmonary nodules had metastases and that some neoplastic diseases, including melanoma and carcinoma of the colon, were more likely to produce solitary metastases.
Considering that even solitary metastatic lung tumors present a varied appearance on TSCT, localized GGA cancer that is attributed to metastasis is expected to be distinctly uncommon. Therefore, localized GGA cancer is not always primary lung adenocarcinoma; however, we still insist on the strong association of localized GGA cancer with lung adenocarcinoma in terms of its high prevalence among neoplastic diseases showing localized GGA on TSCT.
In conclusion, localized GGA cancer, for which tumor size and GGA ratio are presurgical prognostic factors, represents early-stage lung adenocarcinoma in nonsmokers. Further studies are required to validate our results and to better clarify the features of localized GGA cancer as a distinct clinicoradiological entity.
REFERENCES
Tsuchiya R
Implication of the CT characteristics of subcentimeter pulmonary nodules.
Focal area of ground-glass opacity and ground-glass opacity predominance on thin-section CT: discrimination between neoplastic and non-neoplastic lesions.
Proportion of ground-glass opacity on high-resolution computed tomography in clinical T1 N0 M0 adenocarcinoma of the lung: a predictor of lymph node metastasis.
Pathology and Genetics of Tumours of the Lung, Pleura, Thymus and Heart (World Health Organization Classification of Tumours). IARC press,
Lyon2004: 9-124
Supported by a Grant-in-Aid for Young Scientists (Category: B) from the Ministry of Health, Labor and Welfare of Japan and supported by a Grant-in-Aid for the Third-Term Comprehensive 10-Year Strategy for Cancer Control (Category: Japanese General Screening Study for Asbestos-Related Diseases) from the Ministry of Health, Labor, and Welfare of Japan.
Disclosure: The authors declare no conflict of interest.
31457-X&id=gr1.jpg){kind=link}
31457-X&id=gr2.jpg){kind=link}
31457-X&id=gr3.jpg){kind=link}