Biology of colorectal pulmonary metastasis: implications for surgical resection

Abstract

In colorectal cancer, little high grade evidence for cure, life extension, disease modification or palliation achieved by pulmonary metastasectomy exists. This has prompted the pulmonary metastasectomy in colorectal cancer (PulMiCC) trial. Reappraisal of the biological facts on colorectal metastasis may, however, shed light on an alternative avenue of clinical management. Early onset of metastasis, short doubling time and a short disease-free interval are all associated with poor clinical outcomes. Selecting who will be cured (i.e. no occult metastasis) remains the holy grail for pulmonary metastasectomy surgery. Serial CT scans can be utilized to calculate the tumour doubling time by volumetric analysis. Knowing the doubling time and size of the largest metastasis, which by definition is the first cell that has successfully spread from the primary site, the time of initial metastasis can be predicted. More importantly, using the doubling time, calculating the time interval from the primary surgery to the point at which all pulmonary metastases are visible should be possible. Perhaps watchful waiting, with interval CT scanning, followed by pulmonary metastasectomy should be utilized, rather than clinical opinion or randomization in a trial based upon first presentation.

BACKGROUND

In colorectal cancer little high grade evidence for cure, life extension, disease modification or palliation achieved by pulmonary metastasectomy exists. This has prompted the pulmonary metastasectomy in colorectal cancer (PulMiCC) trial [1].

Reappraisal of the biological facts on colorectal metastasis may, however, shed light on an alternative avenue of clinical management. Early onset of metastasis [2], short doubling time [3] and a short disease-free interval [4] are all associated with poor clinical outcomes. With the addition of an interval CT scan patients’ outcomes may be better understood, interpreted and subsequently managed.

Unlike primary carcinoma of the lung, pulmonary metastases from colorectal carcinomas are not thought to metastasize themselves. This does not, however, mean that they do not invade neighbouring structures if left untreated. Even though opinion is divided over resecting pulmonary metastases, data exist to show that the type of cancer, disease-free interval from primary treatment, number of metastases and whether there is unilateral or bilateral metastasis dictate clinical outcomes [1].

Surgical excision of pulmonary metastasis is technically easy and usually involves a wedge resection. Wedge resections as opposed to lobectomy (which is the gold standard for primary lung cancer) are utilized due to the differing tumour biology and lack of metastatic potential of already formed metastasis. Despite being technically easy, morbidity is common and occasionally death secondary to surgery does occur.

Selecting who will be cured (i.e. no occult metastasis) remains the holy grail for pulmonary metastasectomy surgery, hence the PulMiCC trial. Metastasis can form at any time up to and during primary surgical treatment of the colorectal carcinoma, Fig. 1 point B, but not afterwards. The time between points B and E determines the time to wait to assess the number of pulmonary metastases that will be present after primary surgery.

Figure 1:

Demonstrating the use of tumour doubling time. Time points A–F. (A) Time of the firstmetastasis. (B) Surgery on the primary site. The last time a metastasis can occur. (C) The earliest point the first metastasis can be detected (5 mm diameter by CT). (D) A 3-month wait to determine the doubling time. (E) The time point that the last possible metastasis will reach the threshold for CT detection.

HYPOTHESIS

Metastasis is by definition impossible after the tumour has been removed, assuming a local cure. Metastasis can occur due to single cell spread. Clonal expansion then occurs at the site of seeding. The rate of clonal expansion is determined by the tumour doubling time. If serial CT scans are available, the metastasis doubling time can be calculated for lesions that are above the resolution of the CT scanner, on sequential scans. This can be calculated utilizing volumetric analysis. Knowing the doubling time and size of the largest metastasis, which by definition is the first cell that has successfully spread from the primary, the time of initial metastasis can be predicted. The earlier a tumour metastasizes, potentially the poorer is the prognosis. More importantly, knowing the doubling time and the time of primary surgery would then allow calculating the time interval from primary surgery for all pulmonary metastases to be visible on a CT scan, i.e. the time from primary surgery for a single cell to grow large enough to be detected by the resolution of CT scanning.

MODELLING TUMOUR BIOLOGY

The doubling time of colorectal metastasis varies from 30 to 120 days [5]. This means that, to be sure of no metastases at follow-up, a period of between 810 and 3240 days after primary surgery would be necessary, see Fig. 2. It should be noted that one may have to operate in the ‘window of uncertainty’ due to the possibility of the pulmonary metastasis invading a vital intra-thoracic structure in combination with a long doubling time.

Figure 2:

The time from primary surgery to last metastasis detection based on a CT resolution of 5 mm, for various doubling times.

Knowing the tumour doubling time and the time of the first metastasis would enable a certainty of cure to be hypothesized, based on three assumptions.

1. The certainty of cure at the time of the first metastasis, prior to treatment of the primary site, is close to zero.

2. The certainty of cure if a sufficient time has elapsed from primary surgery to metastasectomy, such that no more metastases are possible due to their doubling time, is close to 100%.

3. If a linear time relationship of metastatic spread exists between the first metastasis and primary surgery (no studies to support or refute this assumption), then a certainty score can be postulated as follows:

   

This concept is graphically represented in Fig. 3.

Figure 3:

Predicted certainty of cure based on the doubling time and the time from the first metastasis.

DISCUSSION

Unlike primary lung cancer,  colorectal pulmonary metastases do not spread; they can be followed up until invasion of surrounding structures is imminent. Patients with colorectal metastasis tend to be younger and fitter than patients with lung cancer. Watching a lesion grow to check if other metastases, that are currently below the resolution of CT scanning at initial presentation are present, may mean a lobectomy is necessary. Even in high-risk primary lung cancer patients, with chronic obstructive pulmonary disease, this is only associated with a 1–2% mortality, but due to the interval period of observation, unnecessary metastasectomy may be eliminated in some individuals.

Introducing an interval CT scan to assess the doubling time would help predict the time of initial metastasis and doubling time and establish the certainty that no further metastasis is present. Knowing these data may help us understand who should or should not undergo pulmonary metastasectomy. Although a major step forwards, PulMiCC may not provide an answer as the above basic biological issues are not part of the study.

Metastatic tumour doubling time has previously been shown to be an independent predictor of recurrence after resection of metastases from colorectal cancer in the liver and the lung [6, 7]. A doubling time of less than 45–100 days has been shown to be of poor prognostic significance [3].

The doubling time data from repeat CT scans may also be useful with regard to adjuvant chemotherapy [8]. This may be utilized in conjunction with the carcinoembryonic antigen (CEA) doubling time, which has been shown to be of prognostic influence [9].

The prediction of long-term survival based on metastasectomy surgery with adjuvant therapy clearly being a complicated nonlinear matter, may lend itself to the technique of neural network analysis [10]. Neural networks may be advantageous as they allow multiple nonlinear predictor factor interactions via the hidden layers in the network, unlike the standard statistical tests that only allow a linear interaction between predictor factors. This concept may be further advanced via the use of fuzzy neuronal networks [11], however, clinical validation is needed for all the possible techniques available.

Perhaps watchful waiting, with interval CT scanning, followed by pulmonary metastasectomy should be utilized. More extensive surgery, lobectomy, as opposed to wedge resections, may improve results, but only by operating on fewer patients due to the selection by interval CT scanning.

LIMITATIONS

The tumour doubling time will vary depending on the efficacy of chemotherapy, immunological processes and nutritional changes. This can be modelled for, but would be more complicated, as the doubling times for the metastases during and after chemotherapy would need to be known. The rate of metastasis from a primary colorectal carcinoma is not known, although numerous risk factors for its occurrence are known. Tumour vascular supply is known to be an important risk factor for rate of growth and metastatic potential; however, its quantification and effect on the doubling time is difficult to quantify. Between 5 mm and 2 cm, the size of the majority of metastases, the effect is probably less important than the underlying tumour biology, which is highly variable between individuals [12]. The cost effectiveness and radiation exposure of repeat CT scanning, and the psychological aspects for the patient of waiting will need to be addressed by further work. In this analysis, we have assumed that metastases are all spherical, which is frequently the case; however, a more accurate analysis, especially for non-spherical tumours will require the use of 3D volumetric analysis [13].

Conflict of interest: none declared.