Skip Navigation LinksHome > July 2008 - Volume 3 - Issue 7 > Turning Gray: The Natural History of Lung Cancer Over Time
Journal of Thoracic Oncology:
doi: 10.1097/JTO.0b013e31817c9230
State of the Art: Concise Review

Turning Gray: The Natural History of Lung Cancer Over Time

Detterbeck, Frank C. MD*; Gibson, Christopher J. BA†

Free Access
Article Outline
Collapse Box

Author Information

*Yale Thoracic Oncology Program, Yale Cancer Center, Division of Thoracic Surgery, Department of Surgery and †Medical Student, Yale Medical School, Yale University, New Haven, Connecticut.

Disclosure: The authors declare no conflict of interest.

Dr Frank Detterbeck, M.D., 330 Cedar Street, New Haven, CT 06520-8062. E-mail:

Collapse Box


As newer therapies for lung cancer are being explored it becomes more important to understand the natural history of lung cancer. A systematic review of the data shows that untreated lung cancer is almost uniformly rapidly fatal, even if it is stage I. Analysis of data regarding tumor volume doubling times shows that conventionally detected lung cancers have short mean doubling times, and only a small proportion with very long doubling times. Lung cancers found during the course of a CT screening program have markedly longer mean doubling times and a substantially greater proportion with very long doubling times (>400 days). Models of tumor growth, however, are not understood well enough to use the observed doubling time to predict length of survival without treatment.

Many things are changing in the field of lung cancer, including newer treatments (sublobar resection, radiofrequency ablation, cyberknife, and stereotactic radiosurgery) and different methods of presentation (incidental detection on computed tomography (CT) in an asymptomatic patient).1,2 Results for such interventions have been reported in many nonrandomized studies. The implied argument is generally that the results substantiate the value of the intervention because the outcome is better than if nothing had been done. This calls into question how well we actually know what would have happened if nothing was done to treat a lung cancer. This article seeks to carefully and systematically review the available data regarding the natural history of lung cancer. The primary focus is on early stage (I, II) non-small cell lung cancer (NSCLC), as these patients are the primary focus of most of the newer treatment methods.

Natural history is defined as patient survival in the absence of any active treatment for the cancer. Supportive care measures that palliate specific symptoms are allowed, but not a treatment intended to affect the growth of cancer cells. This may be somewhat ambiguous at times, for example in the case of radiotherapy that is given to shrink a particular focus of tumor blocking an airway or in the brain (while such treatments are given primarily to relieve symptoms, they can significantly affect the length of survival). However, there is general agreement that supportive care measures are permissible when evaluating the natural history of a tumor. Supportive care measures include pain medications, lower dose radiotherapy (<50 Gy) intended to palliate a particular symptom or a surgical procedure intended to palliate a symptom or a local problem. Active treatment is not allowed, such as surgical resection, chemotherapy, and radiotherapy given at higher doses with curative intent.

There are two main ways to assess the natural history. The most obvious is to follow the outcomes of patients who were not treated with any therapeutic intervention (other than basic supportive care). Direct measurement of the natural history is simple but may be difficult to generalize. Those patients who choose or are chosen not to receive any active treatment may be quite different from those receiving treatments, because of comorbidities, psychologic or social issues, all of which have been shown to affect survival.3,4 The differences in the characteristics of treated and untreated patients are perhaps most marked in patients with early stage lung cancer, in whom definition of the natural history is most important.

Another way to estimate survival without treatment is based on observation of the tumor over a period of time, after which an intervention is (usually) undertaken. Estimation of the natural history from a period of observation is probably much more representative of the broad patient population with lung cancer. A period of observation may be more likely to have occurred in patients with slow growing tumors, although it is likely that this selection bias is relatively low. This suppositive is based on the high frequency of missed lesions seen retrospectively on chest radiographs and the frequency of an inherent period of observation of CT-detected lesions (e.g., interval between screening CT and diagnostic CT). However, there is uncertainty about how well the behavior during the period of observation can be extrapolated throughout the life of the tumor.5

Back to Top | Article Outline


This review was prepared by following, as closely as possible, published criteria for a systematic review, even though no widely endorsed standard exists.6,7 The purpose of the project was to define the natural history of untreated lung cancer, and to define growth rates on lung cancer during a period of observation. Specific questions were whether the natural history or growth rates were related to the years included, the histologic cancer type, gender, radiographic appearance, geographic location, the type of study (registry, institutional series etc.) or the method of detection.

The search was designed to find all articles providing data on the natural history and growth rates of NSCLC. We searched three electronic databases (MEDLINE, EMBASE, and the Cochrane Library) for original studies published between 1960 and 2007 that provided data on either the natural history of lung cancer or on the growth rate of untreated NSCLC. We also manually searched the reference sections of all studies identified electronically. The search strategies are available on request from the author. The search was limited to articles published in English. No source of funding was available for this study.

We included articles reporting data on the natural history of untreated patients with NSCLC or on tumor growth rates in the absence of treatment. Both prospective and retrospective study designs were included. Studies providing data on fewer than 20 patients were excluded, as were studies that reported primarily on cases of small-cell lung cancer. No unpublished data were considered. We considered all duplicate publications, but did not list data pertaining to the same cohort of patients twice in the tables (we included cohorts involving some overlapping patients). We excluded articles that reported on only a particular sample or selected patients (e.g., studies using a selected group merely to assess the reliability of measurements of tumor size). No articles were excluded for any other reasons.

Our search strategies yielded 527 studies, which were narrowed to 77 by reviewing the abstracts. These were then reviewed in detail for eligibility for inclusion, and 22 studies were chosen for analysis. In addition, manual search yielded another 13 articles, for a total of 35 that contained data relevant to this study. Primary outcome measures were death, death from lung cancer, and tumor growth. The characteristics of the patients included in the studies were recorded, according to the specific questions to be addressed by our review.

No formal analysis of study quality or publication bias was possible among these retrospective studies. We included data only on patients that met the definitions outlined in this article for natural history, with a few exceptions when it was evident that the outcomes would be similar to those meeting the strict criteria. We included data from studies in which most of patients (≥90%) were untreated. In some instances data was not included in the calculations because the end point or patient cohort was thought to be too dissimilar, but is nevertheless depicted in the tables so the readers can judge the appropriateness of the decision for themselves. In the situations where deviations from the defined end points occurred details of the particular circumstance are provided in the footnotes of the respective table. Quantitative synthesis of the results was limited to simple calculation of averages among the relevant studies. So many factors were poorly described in these retrospective reports that a more sophisticated quantitative analysis appeared inappropriate. Whenever possible we tried to include details of the patient’s characteristics and all of the results for the individual studies so that the reader could assess the consistency of the results and the appropriateness of combining the results for themselves.

Back to Top | Article Outline


Data on the survival of patients with untreated lung cancer is summarized in Tables 1 and 2. These studies have primarily involved lung cancers detected as part of the routine practice of medicine. Many of these patients presented because of symptoms, although some lung cancers were likely found incidentally on a chest radiograph taken for unrelated reasons (especially among patients with stage I tumors). The data demonstrates that survival of untreated patients is distinctively poor. For clinical stages I/II, III, and IV, respectively, the median survival time (MST) is 10, 5, and 3 months; the 1-year survival 39, 17, and 9%; and the 5-year survival only 2%, 0, and 0. Although there is a slight trend to better long-term survival of patients with earlier stages of disease, the effect is not pronounced, is related to lead time (at least in part) and long-term survivors are rare. A small trend can perhaps be perceived towards slightly better survival in more recent series among stage I/II and III patients compared with older patient cohorts, but the effect is marginal. The survival is shown graphically in Figure 1 from one of the largest series, involving almost 23,000 patients from the California cancer registry.

Table 1
Table 1
Image Tools
Table 2
Table 2
Image Tools
Figure 1
Figure 1
Image Tools

In older series, dating back 40 and 50 years, it seems to have been more acceptable to observe patients with lung cancer without administering any active treatment. Therefore, it may be that the patients in older series are more representative of patients in general. On the other hand, the relevance of such old data to the current situation is questionable. Besides the studies cited in Table 1 that reported specifically on untreated patients, a comparison group is available from studies of patients with advanced NSCLC who were randomized to either chemotherapy or basic supportive care during the last 20 years. Two systematic reviews of such studies revealed little difference in MST (3.7 and 4.1 months) compared with the data in Table 1.8,9 Taken together, the data among predominantly stage IV patients suggests a slight trend to better survival in more recent times (a gain of approximately 2 weeks for each decade since the 1960s).

Series of patients with early stage NSCLC who are not treated, suffer from the criticism that these patients are highly selected by negative factors and should therefore be expected to exhibit poor survival. Countering this is the observation that the majority of patients, in fact, died of lung cancer in those series in which this data was recorded. Although the suspicion has been raised that there is an inherent bias to attribute any death in a cancer patient to the cancer (sticky diagnosis or attribution bias), formal analysis has demonstrated that this is relatively infrequent.10–12 Furthermore, this is probably even less prevalent in patients with severe comorbidities, in whom the comorbidities seemed to overshadow the importance of treating a lung cancer (particularly for stage I).

A trend to better survival is suggested among patients with lung cancer that participated in mass screening programs using chest radiographs (Table 2). This is based primarily on one study from Japan, which found a MST of 25 months and a 5-year overall survival of 14%.13 This is corroborated by another study from Japan, which, however, included patients who received a variety of nonsurgical treatments (none of which significantly affected survival).14 A third study from the Memorial Sloan Kettering chest radiograph (CXR) and sputum screening project conducted a decade earlier reported only disease free survival.15 Although it is more difficult to compare, it seems that short- and intermediate-term survival was similar, whereas 5-year survival was worse than in the Japanese series. Furthermore, the patient cohort is not “clean” in that 24% received radiotherapy, although it is not clear at what dose and whether this was intended as curative or palliative treatment. Nevertheless, when all of the data is viewed as a whole, there is at least a suggestion that the natural history of lung cancer detected as part of a CXR screening program may be better than for patients presenting in a more routine fashion). Of course, this apparently longer natural history in screen detected cases includes the effects of lead time bias, and does not necessarily imply a difference in tumor biology. No data is available on the natural history of patients with lung cancer detected as a result of mass screening using CT.

Back to Top | Article Outline


Mean Tumor Doubling Time

Several studies have retrospectively analyzed previous imaging studies to measure tumor volume doubling time (VDT). These are summarized in Table 3. The mean VDT is approximately 135 days for patients diagnosed during routine medical care, 150 days in screening studies involving chest radiographs, and 480 days in screening studies involving CT (Figure 2). This data is fairly consistent among studies involving a particular method of detection (routine care, CXR or CT screening), and thus is quite convincing that CT screening in particular identifies a different patient group. No real trend is apparent in more recent studies compared with older series. The VDT does not seem to be influenced by the reason for the interval between imaging studies, the length of the interval, the years involved or the geographic location of the study.

Table 3
Table 3
Image Tools
Figure 2
Figure 2
Image Tools

A more detailed analysis reveals that women consistently have longer mean doubling times than men by a factor of approximately 2 (range 1.4–2.9).16–19 This gender difference seems to hold true within each method of detection (VDT of approximately 250 versus 125 days in women versus men with detection by CXR screening and 600 versus 350 days in CT screen detected, respectively). A survival advantage for women over men was seen for each cell type except squamous carcinoma in a large SEER database study of treated and untreated patients more than 65-year-old with cI, II NSCLC.20 This suggests that the better survival in women is not merely because of a greater proportion of adenocarcinoma or bronchoalveolar carcinoma (BAC), but reflects a gender-specific difference in biological behavior.

Two studies reported progressively longer VDT for solid, semisolid, and ground glass opacity (GGO) lesions detected by means of CT screening (approximately 150, 475, and 850 days respectively).16,18 However, another study found little difference among these radiographic categories (approximately 500, 570, and 470 days, respectively, p = NS).17

Several studies (Table 4) have consistently documented progressively longer mean VDT by histologic types: squamous carcinoma < adenocarcinoma < BAC < atypical alveolar hyperplasia. This progressive trend seems to be consistent within several methods of detection. The trend to a longer mean VDT is not limited only to BAC, but is seen with adenocarcinoma as well. However, this data is of limited use to the clinician because the histologic type is generally not known until after resection has been accomplished. It is noteworthy that CT screen detected cases as compared with routinely detected cases have a markedly longer VDT for BAC and adenocarcinoma, but there is little difference in patients with squamous carcinoma. It is possible that CT screening has little impact on tumors arising in the airways, (i.e., squamous carcinoma) because of poor visualization of these areas on CT.

Table 4
Table 4
Image Tools
Back to Top | Article Outline
Distribution of Tumor Doubling Times

The data on mean tumor doubling times suggests that the growth rate varies according to the method of detection, and other factors such as gender, histologic type, and possibly the radiographic appearance. These mean values, however, obscure the broad spectrum of doubling times, which are summarized in Table 5. This data clearly demonstrates that there is a broad distribution of tumor doubling times. Furthermore, a marked trend in the distribution of VDT is seen according to the method of detection. In patients undergoing routine care and in CXR screening studies most of tumors have a relatively short VDT (<250 days), whereas in CT screening studies a much broader distribution is seen extending to a very long (>400 days) VDT (Figure 3). The study by Winer-Muram et al., is somewhat of an outlier, with a substantial portion (26%) of routine care detected patients having a VDT of more than 400 days.21 A closer look, however, discloses that the patients in this study were limited to stage I lung cancers seen on a CT scan (although not part of a CT screening program). The decision to exclude this study can be challenged; if it is included the percentages become 41, 42, 11, and 7% for VDT categories of less than 100, 100 to 249, 250 to 399 and ≥400 days, respectively.

Table 5
Table 5
Image Tools
Figure 3
Figure 3
Image Tools

Factors other than the method of detection do not seem to influence the VDT distribution. The time period over which patients were enrolled in the studies does not seem to influence the VDT distribution. Differences exist between lung cancers seen in different parts of the world (e.g., a high proportion of squamous cell cancer in Europe, adenocarcinoma in North America, a higher proportion of BAC and epidermal growth factor receptor (EGFR) mutations in Japan). However, there does not seem to be a correlation of VDT distribution with geographic location.

The proportion of patients with a long VDT (>400 days) seems to be particularly large in women (59 and 37% in two studies) in CT screening studies (versus 36 and 11% in men).17,18 The proportion with long VDT is dramatically increased for GGO lesions with a corresponding decrease in the percent with a short (<100 days) VDT (Figure 4).16–18 In fact approximately 50% of lung cancers that had a GGO appearance had a VDT of more than 800 days (2.2 years) in two CT screening studies.16,18 Semisolid lesions also have a substantial portion (approximately 45%) with long VDT (>400 days), but include some rapidly growing tumors (approximately 10%). In contrast, solid lesions rarely grow slowly (approximately 10%), and many have short VDT (approximately 45%).16–18

Figure 4
Figure 4
Image Tools

There is no evidence that older patients have more slowly growing tumors in the only study providing data regarding this (involving CT screening).18 In fact, the opposite seems to be true, with an increase in the proportion with short (<100 days) VDT in older patients (0, 18, and 36% of patients aged <60, 60–69, and >70, respectively) and arguably a decreasing proportion with long (>400 days) VDT (42, 59, and 18% of patients aged <60, 60–69, and >70, respectively).18 It is unknown how characteristics such as risk factors (e.g., smoking) or willingness to participate in screening study may have influenced these results.

Back to Top | Article Outline
Models of Tumor Growth

To apply the data on observed doubling times to predict the natural history of lung cancer, a review of some concepts of tumor growth is necessary. The two main models of tumor growth are reviewed, and some of the general implications and pitfalls of estimating the natural history.

Back to Top | Article Outline
Exponential Growth

Early laboratory studies included a classic series of elegant experiments involving a murine leukemia model known as L1210.22 In this murine model essentially all of these tumor cells were actively dividing at a constant rate, and could be traced back to a single malignant cell of origin. This led to the widespread adoption of the concept of exponential growth, a model of tumor growth that is still often used today.23 This model fit extremely well with the observed proliferation of leukemia cells in the mouse L1210 model. Early observations of primary lung cancers and metastatic pulmonary nodules on serial CXR were consistent with exponential growth, although the period of observation was limited.24–26

The exponential model postulates that all of the cells are in a growth phase and dividing at a constant rate throughout the life of the tumor. With each doubling time the volume of the tumor doubles; simple mathematics correlates each doubling of volume with an increase in the diameter of approximately 1.26 (the cube root of 2). Assuming the initial tumor cell has a diameter of 10 μm, it would take 30 doublings to reach a diameter of 1 cm, 35 doublings to reach 3 cm, and 40 doublings to reach a diameter of 10 cm (and a weight of ∼1 kg). A tumor of this size in the chest is likely to cause significant life-threatening effects, and therefore the usually accepted time of death is estimated to be at about 40 to 41 doubling times (Figure 5).

Figure 5
Figure 5
Image Tools
Back to Top | Article Outline
Gompertzian Growth Model

Another model, known as the Gompertzian model, stems from an actuarial model that has been applied to tumor growth. This model also assumes that rate of tumor cells division is constant, which has been demonstrated to be the case.27 However, the fraction of tumor cells that is in a growth phase decreases over time. This might be due to the tumor outstripping its blood and nutrient supply, excess excretory products, contact inhibition, or simple crowding.28 This model predicts a progressively longer VDT as the tumor size increases. Mathematically the Gompertz equation states that tumor growth at each instant is exponential, but with a growth constant (growth fraction) that is simultaneously decreasing exponentially.28

Clinical observations over longer periods have shown that most human tumors follow a Gompertzian growth curve, in which growth slows as the tumor gets larger (with the possible exception of Burkitt’s lymphoma which follows a more exponential growth curve).27,29 In fact, Gompertzian growth is also observed in most animal models (Figure 6).28,30,31 The maximum growth rate occurs when the tumor is approximately one-third of its maximal size.27 The Gompertzian model predicts a shorter period of preclinical growth than the exponential model, and longer survival after diagnosis.29

Figure 6
Figure 6
Image Tools
Back to Top | Article Outline
Phases in the Life of a Cancer

It is often assumed that a cancer starts with malignant transformation of a single cell. This cell grows, but has a prolonged period during which the developing tumor is so small it is undetectable by all means. At some point, the tumor is sufficiently large to be detectable if an imaging study is done (i.e., CT or positron emission tomography), but it is still too small to cause any clinical symptoms. Eventually a change in the patient’s well-being prompts medical attention and a diagnosis is made. The time from the initial malignant transformation to clinical detection is the preclinical phase, and generally is at least ¾ of the entire life of the tumor. If a diagnosis is made by an intercurrent imaging study before the onset of symptoms, the time between this incidental diagnosis and the usual diagnosis at clinical presentation represents the lead time. The time between the diagnosis and the patient’s death is the survival time. Obviously, if diagnosis is made earlier, the survival time will be longer (even without any active treatment) because of the lead time.

On the other hand, malignant tumors lead to death not only because of growth of the primary tumor mass, but because of the capacity of tumors to metastasize to distant sites. The historical concept has been abandoned that distant metastases only occurs when the local lymphatics have become overwhelmed and can no longer contain the tumor. The question “When do tumor cells reach the bloodstream? ” has given way to the question “What gives a circulating tumor cell the ability to successfully implant and grow? ” Sensitive analyses have shown that circulating tumor cells are common, even in early stage cancers. For example, circulating tumor cells were detected by reverse transcription-polymerase chain reaction in 40% of 71 patients with NSCLC and 30% of 15 patients with small cell lung cancer patients.32 The presence of circulating tumor cells was unrelated to tumor stage. An acquired genetic change in the tumor cells (i.e., an angiogenesis factor) may be what determines how and when circulating cells develop the ability to form actively growing metastases. Alternatively, it may be changes in the host immune system or microenvironment (i.e., chemokines) that are the key. An understanding of these factors has implications on the ability to predict natural history from observation of the tumor doubling time during a limited period.

Back to Top | Article Outline
Predictions on the Natural History from Doubling Times

It is possible, using the tumor diameter and the observed doubling time, to calculate for an individual patient what the natural history would be if the tumor followed an exponential growth pattern. One study has directly compared actual survival in individual patients with that predicted from the observed doubling time, and found remarkable correlation among nonresected patients with survival time as predicted by the exponential model (r = 0.88, p < 0.01).33 These were patients diagnosed with lung cancer during routine medical care (no screening involved) in whom prior chest radiographs were available. It is not clear why there was a delay in diagnosis (or treatment). Surgical resection resulted in markedly better survival than predicted by the model, but other treatments (given to 24 of 26 nonresected patients) seemed to have little benefit, which is perhaps not surprising given the state of affairs around 1980. Thus this study provides fairly direct data in support of the exponential growth model, although the data is not “clean” (i.e., involving truly untreated patients).

One can attempt to validate growth models in general terms by comparing predictions based on average tumor sizes, average doubling times with average observed survival. Studies have reported an average tumor diameter of 33 mm among cancers detected during routine care,33–37 30 mm in patients undergoing CXR screening,38–40 and 16 mm in patients undergoing CT screening (all scans).1,16,17,37,38,41–45 For patients detected purely by an incidence scan during the course of CT screening the average size is approximately 13 mm.1,42,44

Table 6 tabulates results of calculations based on reported average tumor sizes and doubling times according to the method of detection using the exponential growth model. The calculated survival time based on mean size and mean doubling time is approximately 2.2, 2.7, and 12 years for routine, CXR-screen and CT-screen detected NSCLC, respectively. This is markedly longer than what has been reported in Tables 1 and 2 for routine detection cases, although it is similar for stage I CXR screen detected cases. Furthermore, the estimate of the entire life of the tumor is 15, 17, and 54 years. Even 15 years is difficult to believe, and 54 years seems practically impossible. Thus these estimates call into question the validity of predictions about the survival time based on observed doubling times and the exponential model (including the assumption that death occurs on average after 40–41 doublings).

Table 6
Table 6
Image Tools

Predicting growth over the entire life of a tumor using the Gompertzian model is more complex. This is in part because the doubling time varies depending upon the size of the tumor. Thus the applicability of the observed doubling time must be estimated by taking into account the tumor size during the period of observation. In general terms, however, the Gompertzian model would predict that the growth rate before clinical detection would be more rapid than the exponential model, and slower afterward. Thus the survival times after clinical detection would likely be two to three times longer than predicted by the exponential model. These predictions do not fit the observed survival times for lung cancers detected by routine care or by CXR screening (no observations exist with which to correlate survival of CT detected lung cancers).

Back to Top | Article Outline


The observed survival of untreated patients in whom lung cancer is detected as part of routine care is uniformly poor, suggesting that these patients should be managed aggressively and expeditiously. The median survival of patients with routinely detected clinical stage I, II tumors is only about 10 months if untreated. A brief trial of antibiotics to differentiate between infection and malignancy may sometimes be justified. However, in many instances such a “diagnostic delay” is not justified because the radiographic appearance so clearly suggests lung cancer and not infection (discrete, solid, spiculated mass). Unless the radiographic appearance is strongly suggestive of scar, a period of delay to observe growth does not seem to be warranted. Furthermore, a prolonged sequential work-up involving a battery of imaging tests and biopsy procedures often takes many weeks to complete. Prompt referral to a specialist appears to result in a more rapid and streamlined evaluation and initiation of treatment. This has led the relevant medical organizations in several countries (Great Britain, Canada, France, Sweden) to either mandate or strongly recommend through guidelines that patients with a suspicion of lung cancer be referred and seen promptly (within 1–2 weeks) by a thoracic specialist, in general in conjunction with a multidisciplinary team. However, the amount of benefit that is realized from rapid evaluation and initiation of treatment is unclear.46 It is suggested that the patient be seen in an organized multidisciplinary program,47,48 because of data suggesting that this promotes the delivery of evidence-based care and improves outcomes.49–53

The data regarding the distribution of tumor doubling times shows that lung cancers involve a wide spectrum of growth rates. A broader appreciation of this fact is warranted, because CT imaging in particular appears to affect the spectrum of disease that is encountered. The presence of CT screening programs and the increased prevalence of CT imaging (outside of a CT screening program) makes this observation particularly important at this time. Direct natural history data to corroborate this observation, however, is not available. Nevertheless, the increasingly broad spectrum of disease leads to speculation that with a better understanding it may be possible to match the aggressiveness of the treatment to the aggressiveness of the particular tumor. Obviously, this will require not only methods to define average characteristics of types of tumors, but confidence that one can define the aggressiveness of the tumor of an individual patient.

The spectrum of lung cancer detected as a result of CT screening is markedly different than that of patient diagnosed during routine medical care. Tumors found in CT screening programs have markedly longer doubling times, and a substantially greater proportion with very long (>400 days) doubling times. There is little difference between routine care detected tumors and CXR screen detected tumors, although a slight trend is apparent. Surprisingly, there is little difference in routine care detected cancers currently compared with those detected decades ago. A minor trend towards less rapidly growing tumors in more recent times is suggested, perhaps related to a greater prevalence of imaging for individuals in general.

Several important implications follow from the observations that a spectrum of aggressiveness exists among lung cancers and that the proportion of patients along the spectrum is affected by the method of detection, and other characteristics such as the radiographic appearance (GGO versus semisolid versus solid). First, it is probably inappropriate to use data generated from routinely detected patients to make estimates about the outcomes of CT screen detected patients. This applies to an estimate of a potential mortality benefit as well as an assessment of cost-effectiveness. Similarly, we must be careful in the interpretation of phase II studies of novel treatment approaches such as Radiofrequency Ablation and Stereotactic radiosurgery. Second, evaluation of treatments for less aggressive tumors may require a prolonged period of follow-up before the true impact of the treatment can be assessed. Finally, the greater prevalence of indolent tumors raises the question that over-treatment and over-diagnosis may be a substantial risk in some settings.

Unfortunately, models of tumor growth do not seem to correlate well with estimations derived from observed doubling times. Thus predictions about the natural history made from doubling times are not currently accurate enough to justify not treating a diagnosed lung cancer in most clinical situations. Although observation alone may well turn out to be appropriate for some patients, this must be studied further. This conclusion is further strengthened by the lack of understanding of the relationship between (indolent) tumor growth and the development of distant metastases.

A particularly poignant issue is the management of patients with a GGO. The data regarding VDT suggests that this group has a high proportion of slow growing tumors (40–90% with VDT of >400 days among CT screen detected tumors). Several factors contribute to the difficulty in defining the appropriate management of a GGO. The correlation between radiographic characteristics and histologic diagnosis (atypical alveolar hyperplasia, BAC, or well differentiated adenocarcinoma) is unclear.54–56 Furthermore, in a substantial portion of patients with a GGO the presence of a malignancy is associated with an increase in the density of the lesion without growth, or even with shrinkage of the lesion.17,57–59 Thus a GGO may represent a benign lesion or a cancer, and growth alone does not always differentiate these.

The data cited in this review stems largely from formal screening studies, and an important question is how well this applies to patients who are not participating in a screening program, given the rapidly increasing number of CT scans that are done for unrelated reasons.60 Recent studies among treated patients have suggested the same good prognosis among patients with a CXR screen detected lung cancer and one detected incidentally by a CXR done for other reasons during work-up of an unrelated medical condition—in contrast to a lung cancer detected by routine clinical presentation.61 This observation held up even among stage pI tumors, suggesting that the method of detection, whether part of a formal screening program or not, selected patients with a different biological behavior.61 No comparison is available between lung cancers detected incidentally by a CT scan and one detected in a CT screening program.

Many questions remain about the natural history of lung cancer, and insufficient and conflicting data exists in many areas. Nevertheless, two main concepts clearly emerge. First, untreated lung cancer as detected during routine clinical care is rapidly fatal, even among patients with stage I NSCLC. However, a second central concept is that the spectrum of growth rates of lung cancer seems to be changing significantly. The evidence is strong and consistent that CT screening identifies a somewhat different cohort of patients with lung cancer that includes many more patients with slow growing tumors. This phenomenon may also extend to GGOs that are discovered incidentally because of an increasing frequency of imaging in general. Thus our changing health care environment demands that we view lung cancer as a spectrum of shades of gray. It is important that we develop a thorough appreciation of the range of disease behavior, because it is crucial to appropriately evaluate data regarding practically every aspect of lung cancer—including screening, diagnosis, staging, and treatment.

Back to Top | Article Outline


Review of the data on the natural history of lung cancer leads to the following conclusions:

1. Lung cancer as diagnosed during routine medical care is rapidly fatal, even among clinical stage I tumors. Patients who are found to have a lesion suspicious for lung cancer should be referred promptly to a specialist for evaluation. A period of observation by the internist or family practitioner is not justified other than possibly for a brief period to exclude pneumonia. There is only a minimal trend to better survival in more recent years.

2. There is a wide variation in tumor growth rates. Lung cancer is a heterogeneous disease with a spectrum of growth characteristics.

3. The proportion of slowly growing tumors VDT (>400 days) is substantially increased in patients diagnosed as part of screening program (especially CT screening). It is inappropriate to apply data regarding the natural history of lung cancer from populations of patients detected during routine care to patients detected as part of a CT screening program.

4. The larger proportion of slowly growing tumors in CT screening programs suggests that over-treatment and over-diagnosis may exist.

5. Limited data suggests that the increase in indolent tumors seen in CT screening programs may also be present among tumors found by incidental (nonscreening) CT imaging.

6. Models of tumor growth are not sufficiently clear to allow accurate prediction of the natural course from knowledge of the VDT of the tumor.

7. Awareness of the changing spectrum of aggressiveness among lung cancers is needed to correctly interpret the results of non-randomized studies.

Back to Top | Article Outline


1. Henschke C, Yankelevitz D, Libby DM, et al. Survival of patients with stage I lung cancer detected on CT screening. N Engl J Med 2006;355:1763–1771.

2. Onishi H, Nagata Y, Shirato H, et al. Stereotactic hypofractionated high-dose irradiation for stage I non-small cell lung carcinoma: clinical outcomes in 273 cases of a Japanese multi-institutional study. ASCO 2004;22; abstract no. 7003.

3. Yellen SB, Cella DF. Someone to live for: social well-being, parenthood status, and decision-making in oncology. J Clin Oncol 1995;13:1255–1264.

4. Battafarano RJ, Piccirillo JF, Meyers BF, et al. Impact of comorbidity on survival after surgical resection in patients with stage I non-small cell lung cancer. J Thorac Cardiovasc Surg 2002;123:280–287.

5. Stinchcombe TE, Socinski MA. General aspects of chemotherapy. In Detterbeck FC, Rivera MP, Socinski MA, Rosenman JG, (Eds.), Diagnosis and Treatment of Lung Cancer: An Evidence-Based Guide for the Practicing Clinician. Philadelphia, PA: WB Saunders, 2001. Pp. 162–173.

6. Weed DL. Methodologic guidelines for review papers. J Natl Cancer Inst 1997;89:6–7.

7. Moher D, Tetzlaff J, Tricco AC, Sampson M, Altman DG. Epidemiology and reporting characteristics of systematic reviews. PLoS Med 2007;4:e78.

8. Socinski MA, Morris DE, Masters GA, Lilenbaum R. Chemotherapeutic management of stage IV non-small cell lung cancer. Chest 2003;123:226S–243S.

9. Socinski MA, Crowell R, Hensing TE, et al. Treatment of non-small cell lung cancer, stage IV: ACCP evidence-based clinical practice guidelines (2nd edition). Chest 2007;132:277S–289S.

10. Satariano W, Ragland K, Van Den Eeden S. Cause of death in men diagnosed with prostate carcinoma. Cancer 1998;83:1180–1188.

11. Penson DF, Albertsen PC, Nelson PS, Barry M, Stanford JL. Determining cause of death in prostate cancer: are death certificates valid? J Natl Cancer Inst 2001;93:1822–1823.

12. Albertsen PC, Walters S, Hanley JA. A comparison of cause of death determination in men previously diagnosed with prostate cancer who died in 1985 or 1995. J Urol 2000;163:519–523.

13. Sobue T, Suzuki T, Matsuda M, et al. Survival for clinical stage I lung cancer not surgically treated: comparison between screen-detected and symptom-detected cases. Cancer 1992;69:685–692.

14. Motohiro A, Ueda H, Komatsu H, Yanai N, Mori T. Prognosis of non-surgically treated, clinical stage I lung cancer patients in Japan. Lung Cancer 2002;36:65–69.

15. Flehinger BJ, Kimmel M, Melamed MR. The effect of surgical treatment on survival from early lung cancer: implications for screening. Chest 1992;101:1013–1018.

16. Hasegawa M, Sone S, Takashima S, et al. Growth rate of small lung cancers detected on mass CT screening. Br J Radiol 2000;73:1252–1259.

17. Lindell RM, Hartman TE, Swensen SJ, et al. Five-year lung cancer screening experience: CT appearance, growth rate, location, and histologic features of 61 lung cancers. Radiology 2007;242:555–562.

18. Sone S, Nakayama T, Honda T, et al. Long-term follow-up study of a population-based 1996–1998 mass screening programme for lung cancer using mobile low-dose spiral computed tomography. Lung cancer 2007;58:329–341.

19. Usuda K, Saito Y, Sagawa M, et al. Tumor doubling time and prognostic assessment of patients with primary lung cancer. Cancer 1994;74:2239–2244.

20. Wisnivesky JP, Halm EA. Sex differences in lung cancer survival: do tumors behave differently in elderly women? J Clin Oncol 2007;25:1705–1712.

21. Winer-Muram HT, Jennings SG, Tarver RD, et al. Volumetric growth rate of stage I lung cancer prior to treatment: serial CT scanning. Radiology 2002;223:798–805.

22. Skipper HE, Schabel FM Jr, Wilcox WS. On the criteria and kinetics associated with “curability” of experimental leukemia. Cancer Chemother Rep 1964;35:1–111.

23. Bach PB, Silvestri G, Hanger M, Jett JR. Screening for lung cancer: ACCP evidence-based clincial practice guidelines, (2nd edition). Chest 2007;132:69S–77S.

24. Steele JD, Buell P. Asymptomatic solitary pulmonary nodules: host survival, tumor size, and growth rate. J Thorac Cardiovasc Surg 1973;65:140–151.

25. Spratt J, Spratt T. Rates of growth of pulmonary metastases and host survival. Ann Surg 1964;159:161–171.

26. Weiss W. Peripheral measurable bronchogenic carcinoma. Growth rate and period of risk after therapy. Am Rev Respir Dis 1971;103:198–208.

27. DeVita VT Jr. Principles of cancer management: chemotherapy. In: DeVita VT Jr, Hellman S, Rosenberg SA, (Eds.), Cancer: Principles & Practice of Oncology, 5th Ed. Philadelphia: Lippincott-Raven, 1997. Pp. 333–347.

28. Schabel FM Jr. Concepts for systemic treatment of micrometastases. Cancer 1975;35:15–24.

29. Norton L. A Gompertzian model of human breast cancer growth. Cancer Res 1988;48:7067–7071.

30. Gunduz N, Fisher B, Saffer EA. Effect of surgical removal on the growth and kinetics of residual tumor. Cancer Res 1979;39:3861–3865.

31. Simpson-Herrin L, Sanford AH, Holmquist JP. Effects of surgery in the cell kinetics of residual tumor. Cancer Treat Rep 1976;60:1744–1760.

32. Peck K, Sher Y-P, Shih J-Y, et al. Detection and quantitation of circulating cancer cells in the peripheral blood of lung cancer patients. Cancer Res 1998;58:2761–2765.

33. Mizuno T, Masaoka A, Ichimura H, et al. Comparison of actual survivorship after treatment with survivorship predicted by actual tumor-volume doubling time from tumor diameter at first observation. Cancer 1984;53:2716–2720.

34. Geddes DM. The natural history of lung cancer: a review based on rates of tumour growth. Br J Dis Chest 1979;73:1–17.

35. McGarry RC, Song G, des Rosiers P, Timmerman R. Observation-only management of early stage, medically inoperable lung cancer: poor outcome. Chest 2002;121:1155–1158.

36. Raz DJ, Zell JA, Ou SHI, et al. Natural history of stage I non-small cell lung cancer: implications for early detection. Chest 2007;132:193–199.

37. Altorki N, Kent M, Pasmantier M. Detection of early-stage lung cancer: computed tomographic scan or chest radiograph? J Thorac Cardiovasc Surg 2001;121:1053–1057.

38. Kashiwabara K, Kohshi S. Outcome in patients with lung cancer invisible on chest roentgenograms but detected only by helical computed tomography. Respirology 2006;11:592–597.

39. Kashiwabara K, Koshi S, Ota K, Tanaka M, Toyonaga M. Outcome in patients with lung cancer found retrospectively to have had evidence of disease on past lung cancer mass screening roentgenograms. Lung Cancer 2002;35:237–241.

40. Kashiwabara K, Koshi S, Itonaga K, et al. Outcome in patients with lung cancer found on lung cancer mass screening roentgenograms, but who did not subsequently consult a doctor. Lung Cancer 2003;40:67–72.

41. Lindell RM, Hartman TE, Swensen SJ, et al. Lung cancer screening experience: a retrospective review of PET in 22 non-small cell lung carcinomas detected on screening chest CT in a high-risk population. Am J Roentgenol 2005;185:126–131.

42. Sobue T, Moriyama N, Kaneko M, et al. Screening for lung cancer with low-dose helical computed tomography: anti-lung cancer association project. J Clin Oncol 2002;20:911–920.

43. Sone S, Takashima S, Li F, et al. Mass screening for lung cancer with mobile spiral computed tomography scanner. Lancet 1998;351:1242–1245.

44. Pastorino U, Bellomi M, Landoni C, et al. Early lung-cancer detection with spiral CT and positron emission tomography in heavy smokers: 2-year results. Lancet 2003;362:593–597.

45. Swensen SJ, Jett JR, Hartman TE, et al. CT screening for lung cancer: five-year prospective experience. Radiology 2005;235:259–265.

46. Spiro SG, Gould MK, Colice GL. Initial evaluation of the patient with lung cancer: symptoms, signs, laboratory tests, and paraneoplastic syndromes: ACCP evidenced-based clinical practice guidelines, (2nd edition). Chest 2007;132:149S–160S.

47. Alberts WM, Bepler G, Hazelton T, Ruckdeschel JC, Williams JH Jr. Lung cancer. Practice organization. Chest 2003;123:332S–337S.

48. BTS. BTS recommendations to respiratory physicians for organising the care of patients with lung cancer. Thorax 1998;53:S1–S8.

49. Hillner BE, Smith TJ, Desch CE. Hospital and physician volume or specialization and outcomes in cancer treatment: importance in quality of cancer care. J Clin Oncol 2000;18:2327–2340.

50. Silvestri GA, Handy J, Lackland D, Corley E, Reed CE. Specialists achieve better outcomes than generalists for lung cancer surgery. Chest 1998;114:675–680.

51. Bach PB, Cramer LD, Schrag D, et al. The influence of hospital volume on survival after resection for lung cancer. N Engl J Med 2001;345:181–188.

52. Fergusson RJ, Thomson CS, Brewster DH, et al. Lung cancer: the importance of seeing a respiratory physician. Eur Respir J 2003;21:606–610.

53. Smith TJ, Hillner BE. Ensuring quality cancer care by the use of clinical practice guidelines and critical pathways. J Clin Oncol 2001;19:2886–2897.

54. Ikeda K, Awai K, Mori T, et al. Differential diagnosis of ground-glass opacity nodules: CT number analysis by three-dimensional computerized quantification. Chest 2007;132:984–990.

55. Kim H, Shim Y, Lee K, et al. Persistent pulmonary nodular ground-glass opacity at thin-section CT: histopathologic comparisons. Radiology 2007;245:267–275.

56. Ohtsuka T, Watanabe K, Kaji M, Naruke T, Suemasu K. A clinicopathological study of resected pulmonary nodules with focal pure ground-glass opacity. Eur J CardioThorac Surg 2006;30:160–163.

57. Takashima S, Maruyama Y, Hasegawa M, et al. CT findings and progression of small peripheral lung neoplasms having a replacement growth pattern. AJR Am J Roentgenol 2003;180:817–826.

58. Kakinuma R, Ohmatsu H, Kaneko M, et al. Progression of focal pure ground-glass opacity detected by low-dose helical computed tomography screening for lung cancer. J Comput Assist Tomogr 2004;28:17–23.

59. Jennings SG, Winer-Muram HT, Tann M, Ying J, Dowdeswell I. Distribution of stage I lung cancer growth rates determined with serial volumetric CT measurements. Radiology 2006;241:554–563.

60. Brenner DJ, Hall EJ. Computed tomography—an increasing source of radiation exposure. N Engl J Med 2007;357:2277–2284.

61. Hanagiri T, Sugio K, Mizukami M, et al. Postoperative prognosis in patients with non-small cell lung cancer according to the method of initial detection. J Thorac Oncol 2007;2:907–911.

62. Hyde L, Yee J, Wilson R, Patno M. Cell type and the natural history of lung cancer. JAMA 1965;193:140–142.

63. Roswit B, Patno ME, Rapp R, et al. The survival of patients with inoperable lung cancer: a large-scale randomized study of radiation therapy versus placebo. Radiology 1968;90:688–697.

64. Hyde L, Wolf J, McCracken S, Yesner R. Natural course of inoperable lung cancer. Chest 1973;64:309–312.

65. Zelen M. Keynote address on biostatistics and data retrieval. Cancer Chemother Rep 3 1973;4:31–42.

66. Paul A, Marelli D, Wilson JAS, Chiu RC, Mulder DS. Does the surgical trauma of “exploratory thoracotomy” affect survival of patients with bronchogenic carcinoma? Can J Surg 1989;32:322–327.

67. Reinfuss M, Skolyszewski J, Kowalska T, Rzepecki W, Kociolek D. Palliative radiotherapy in asymptomatic patients with locally advanced, unresectable, non-small cell lung cancer. Strahlenther Onkol 1993;169:709–715.

68. Vrdoljak E, Mise K, Sapunar D, Rozga A, Marusic M. Survival analysis of untreated patients with non-small-cell lung cancer. Chest 1994;106:1797–1800.

69. Leung WT, Shiu WCT, Pang JCK, et al. Combined chemotherapy and radiotherapy versus best supportive care in the treatment of inoperable non-small-cell lung cancer. Oncology 1992;49:321–326.

70. Chadha A, Ganti A, Sohi J, Sahmoun A, Mehdi SA. Survival in untreated early stage non-small cell lung cancer. Anticancer Res 2005;25:3517–3520.

71. Kyasa MJ, Jazieh AR. Characteristics and outcomes of patients with unresected early-stage non-small cell lung cancer. South Med J 2002;95:1149–1152.

72. Henschke CI, Wisnivesky JP, Yankelevitz DF, Miettinen OS. Small stage I cancers of the lung: genuineness and curability. Lung Cancer 2003;39:327–330.

73. Wisnivesky JP, Bonomi M, Henschke C, Iannuzzi M, McGinn T. Radiation therapy for the treatment of unresected stage I-II non-small cell lung cancer. Chest 2005;128:1461–1467.

74. Garland LH, Coulson W, Wollin E. The rate of growth and apparent duration of untreated primary bronchial carcinoma. Cancer 1963;16:694–707.

75. Arai T, Kuroishi T, Saito Y, et al. Tumor doubling time and prognosis in lung cancer patients: evaluation from chest films and clinical follow-up study. Jpn J Clin Oncol 1994;24:199–204.

76. Yankelevitz DF, Kostis WJ, Henschke CI, et al. Overdiagnosis in chest radiographic screening for lung carcinoma: frequency. Cancer 2003;97:1271–1275.

77. Weiss W, Boucot K, Cooper D. Growth rate in the detection and prognosis of bronchogenic carcinoma. JAMA 1966;198:1246–1252.

78. Kerr K, Lamb D. Actual growth rate and tumour cell proliferation in human pulmonary neoplasms. Br J Cancer 1984;50:343–349.

Cited By:

This article has been cited 15 time(s).

Archives of Pathology & Laboratory Medicine
High-Resolution Computed Tomography Screening for Lung Cancer Unexpected Findings and New Controversies Regarding Adenocarcinogenesis
Chirieac, LR; Flieder, DB
Archives of Pathology & Laboratory Medicine, 134(1): 41-48.

Annals of Thoracic Surgery
Impact of Smoking Cessation Before Resection of Lung Cancer: A Society of Thoracic Surgeons General Thoracic Surgery Database Study
Mason, DP; Subramanian, S; Nowicki, ER; Grab, JD; Murthy, SC; Rice, TW; Blackstone, EH
Annals of Thoracic Surgery, 88(2): 362-371.
Journal of Thoracic Oncology
Time to Diagnosis of Lung Cancer Technical and Pyschological Factors that Slow Down Diagnostic and Treatment Timelines
Pujol, JL; Quantin, X
Journal of Thoracic Oncology, 4(): 1192-1194.

The fruits of our efforts: time for a different view of lung cancer and CT screening
Detterbeck, F
Thorax, 64(6): 465-466.
Lung Cancer
Doubling time of lung cancer determined using three-dimensional volumetric software: Comparison of squamous cell carcinoma and adenocarcinoma
Honda, O; Johkoh, T; Sekiguchi, J; Tomiyama, N; Mihara, N; Sumikawa, H; Inoue, A; Yanagawa, M; Daimon, T; Okumura, M; Nakamura, H
Lung Cancer, 66(2): 211-217.
The New Lung Cancer Staging System
Detterbeck, FC; Boffa, DJ; Tanoue, LT
Chest, 136(1): 260-271.
Lung Cancer
Second primary lung cancer or double cancer of the aerodigestive tract?
Debevec, L
Lung Cancer, 65(2): 255.
Thoracic Surgery Clinics
Systematic Approach to the Management of the Newly Found Nodule on Screening Computed Tomography Role of Dedicated Pulmonary Nodule Clinics
Korst, RJ
Thoracic Surgery Clinics, 23(2): 141-+.
Evolutionary Applications
Cancer as a moving target: understanding the composition and rebound growth kinetics of recurrent tumors
Foo, J; Leder, K; Mumenthaler, SM
Evolutionary Applications, 6(1): 54-69.
Counterpoint: Are Limited Resections Appropriate in Non-small Cell Lung Cancer? No Don't Overdo It, and Don't Get Confused
Detterbeck, FC
Chest, 141(3): 590-592.
Management of Part-Solid Nodules
Bommart, S; Kovacsik, HV; Pujol, JL
Radiology, 268(1): 306.

Clinical Lung Cancer
The Effect of a Lung Cancer Care Coordination Program on Timeliness of Care
Alsamarai, S; Yao, XP; Cain, HC; Chang, BW; Chao, HH; Connery, DM; Deng, YH; Garla, VN; Hunnibell, LS; Kim, AW; Obando, JA; Taylor, C; Tellides, G; Rose, MG
Clinical Lung Cancer, 14(5): 527-534.
Journal of Thoracic Oncology
Anatomy, Biology and Concepts, Pertaining to Lung Cancer Stage Classification
Detterbeck, FC; Tanoue, LT; Boffa, DJ
Journal of Thoracic Oncology, 4(4): 437-443.
PDF (423) | CrossRef
Journal of Thoracic Oncology
A Systematic Review but Systematically Confounded?
Detterbeck, FC
Journal of Thoracic Oncology, 5(5): 754-756.
PDF (553) | CrossRef
Journal of Thoracic Oncology
The International Association for the Study of Lung Cancer Staging Project: Prognostic Factors and Pathologic TNM Stage in Surgically Managed Non-small Cell Lung Cancer
Chansky, K; Sculier, J; Crowley, JJ; Giroux, D; Van Meerbeeck, J; Goldstraw, P; on behalf of the International Staging Committee and Participating Institutions,
Journal of Thoracic Oncology, 4(7): 792-801.
PDF (632) | CrossRef
Back to Top | Article Outline

Natural history; Lung cancer; Doubling time; Screening

© 2008International Association for the Study of Lung Cancer


Article Tools



Search for Similar Articles
You may search for similar articles that contain these same keywords or you may modify the keyword list to augment your search.

Other Ways to Connect



Visit on your smartphone. Scan this code (QR reader app required) with your phone and be taken directly to the site.

 For additional oncology content, visit LWW Oncology Journals on Facebook.